Current Issues in Medieval England 3631862741, 9783631862742, 9783631862957

Citation preview

Current Issues in Medieval England

STUDIES IN ENGLISH MEDIEVAL LANGUAGE AND LITERATURE Edited by Magdalena Bator

Advisory Board: John Anderson (Methoni, Greece), Ulrich Busse (Halle), Isabel de la Cruz-Cabanillas (Alcala, Spain) Olga Fischer (Amsterdam), Marcin Krygier (Poznań), Peter Lucas (Cambridge), Donka Minkova (Los Angeles), Akio Oizumi (Kyoto), Katherine O’Brien O’Keeffe (UC Berkeley, USA), Hans Sauer (Munich), Liliana Sikorska (Poznań), Hans Sauer (Munich), Jeremy Smith (Glasgow), Jerzy Wełna (Warsaw)

Volume 59

Letizia Vezzosi (ed.)

Current Issues in Medieval England

Bibliographic Information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data is available in the internet at http://dnb.d-nb.de. Library of Congress Cataloging-in-Publication Data A CIP catalog record for this book has been applied for at the Library of Congress. This publication was financially supported by the Department of Education, Languages, Interculture, Literatures and Psychology at the University of Florence.

ISSN 1436-7521 ISBN 978-3-631-86274-2 (Print) E-ISBN 978-3-631-86295-7 (E-PDF) E-ISBN 978-3-631-86310-7 (EPUB) DOI 10.3726/b18761 © Peter Lang GmbH Internationaler Verlag der Wissenschaften Berlin 2021 All rights reserved. Peter Lang Edition is an Imprint of Peter Lang GmbH. Peter Lang – Berlin ∙ Bern ∙ Bruxelles ∙ New York ∙ Oxford ∙ Warszawa ∙ Wien All parts of this publication are protected by copyright. Any utilisation outside the strict limits of the copyright law, without the permission of the publisher, is forbidden and liable to prosecution. This applies in particular to reproductions, translations, microfilming, and storage and processing in electronic retrieval systems. This publication has been peer reviewed. www.peterlang.com

Contents Contents ������������������������������������������������������������������������������������������������������������������� 5 List of Contributors �������������������������������������������������������������������������������������������� 7 Preface ����������������������������������������������������������������������������������������������������������������������� 9 Part I: Textual Interlacing Claudio Cataldi A Re-​assessment of Poema Morale and its Influence on Penitence for Wasted Life ���������������������������������������������������������������������������������������������������������������  15 Gabriella Del Lungo Camiciotti –​Julia Bolton Holloway The Soul a City: Margery meets Julian �����������������������������������������������������������������  33 Omar Khalaf Patronage, Print and the Education of the Gentry in Late Medieval England: The Case of Earl Rivers’s Dicts and Sayings of the Philosophers ��������  45 Patrizia Lendinara On the Trail of Bibbesworth ����������������������������������������������������������������������������������  59 Alessandra Petrina The Construction of the European Intellectual: Petrarch, Humanism, and Middle English Literature ��������������������������������������������������������������������������������������  87

Part II: Borrowing and Lexicon Angelika Lutz Changes of Political Rule and the Changing Use of OE Gærsum(a) ‘Treasure’ (< ON Gersemi) in Middle English ���������������������������������������������������  111

6

Contents

Rafał Molencki The Rise of the Verb Happen in Middle English �����������������������������������������������  129 Louise Sylvester –​Harry Parkin –​Richard Ingham Patterns of Borrowing, Obsolescence and Polysemy in the Technical Vocabulary of Middle English �����������������������������������������������������������������������������  143

Part III: Language at Different Levels Martti Mäkinen Rhetorical Re-​analysis of Metadiscourse Items in Henry Daniel’s Middle English Prologue to Liber Uricrisiarum ��������������������������������������������������������������  171 Gjertrud F. Stenbrenden The Evidentiary Status of Back Spellings and English Historical Phonology ��������������������������������������������������������������������������������������������  195 Merja Stenroos What, If Anything, Are Middle English Dialects? Some Thoughts on a Changing Concept ������������������������������������������������������������������������������������������������  217 Brita Wårvik The Demise of Ambiguous Adverb/​Conjunctions and Manuscript Variation: A Case Study of Tho, Then and When in the Chronicle of Robert of Gloucester �����������������������������������������������������������������������������������������������������������  245 Richard Zimmermann The Loss of Negative Concord with Negative PPs ��������������������������������������������  273

Index of terms and names ���������������������������������������������������������������������������  301

List of Contributors Julia Bolton Holloway University of Colorado

Harry Parkin University of Westminster

Claudio Cataldi University of Bristol

Alessandra Petrina University of Padova

Gabriella Del Lungo University of Florence

Rafał Molencki University of Silesia

Omar Khalaf University of Padova

Gjertrud F. Stenbrenden University of Oslo –​Inland Norway University of Applied Sciences

Richard Ingham Birmingham City University Angelika Lutz University of Erlangen Patrizia Lendinara University of Palermo Martti Mäkinen Hankel School of Economics –​Helsinki

Merja Stenroos University of Stavanger Louise Sylvester University of Westminster Brita Wårvik Åbo Akademi University –​Turku Richard Zimmermann University of Manchester

Preface “Middle English […] is a very convenient term denoting […] “the transition period in the history of English between Old and Modern English”. “It was preceded by a period, during which the linguistic changes visible in the thirteenth and the following centuries must have begun to operate (c. 1050–​1150) and was followed by a similar period of transition between Middle and Modern English (c. 1450–​1500). “The term itself might suggest a linguistic entity, which it was not” (Fisiak 1968: 10–​11). At this point I would like to recall Prof. Jacek Fisiak, who launched the series of International Conferences on Middle English and who was unable (for the first time) to attend the one in Florence for reasons of health, and this unorthodox quotation from his A short grammar of Middle English to give an idea of the complexity the term Middle English still entails. An iconic simile for Middle English could be the lenticular mosaic: changing perspective, one sees different images which still communicate and interrelate to one another; looking from afar, one gets a defined and sharp picture which inevitably shatters into a multitude of tiles, i.e. linguistic and literary items, projecting into the past and the future, whose meaning can only be understood if they are seen within their “mosaic frame”. According to the simile, the papers selected in this volume from the 11th International Conference on Middle English (ICOME11), held at the University of Florence from 5 to 8 February 2019, constitute each a single tile of different images in the bigger mosaic of Middle English. One hundred and fifteen scholars from several countries, i.e., from Austria, Belgium, Czech Republic, Finland, France, Germany, Kazakhstan, the Netherlands, Norway, Oman, Poland, Russia, South Korea, Spain, Switzerland, Ukraine, USA and Italy participated in the conference. Sixty-​four papers were presented and thirteen have been selected for publication in the volume. The papers presented at the Florence conference ask new research questions about Medieval England: starting from traditional topics in historical linguistics and literature from different perspectives, extending ongoing debates to new kinds of material and/​or through new tools, i. e. new corpora, and new approaches, questioning traditional categories of inquiry and theory, and detecting the interwoven contribution of each topic to the development of English language and English literature through the network of influences and references in which the relevant linguistic, literary or cultural phenomenon is placed. The volume reflects the scope of these new approaches

10

Preface

and views within the field. The traditional subdivision into “Language” and “Literature” of ICOME volumes is supplemented by a third section “Borrowing and the Lexicon”, which is devoted to one of the most intensively investigated linguistic field in Middle English studies, bridging the literary section, named “Textual Interlacing”, and the linguistic section, “Language at Different Levels”. The first section “Textual Interlacing” includes five papers focussing on the influential relations between texts, personalities and international authors. The opening chapter by Cataldi re-​assesses the debt of influence of Penitence for Wasted Life to Poema Morale in terms of both content and metre through a detailed philological, metrical and linguistic analysis. Similarly, Lendinara’s careful philological study follows the influence of the Tretiz by Walter of Bibbesworth in Middle English and French literature, and multilingual glossaries over the two centuries. The attention of the author shifts from the literary work to its society in Khalaf ’s paper, which investigates the production and circulation of Earl River’s Dicts and Sayings of the Philosophers as part of a precise cultural plan of dissemination of moral and educational literature outside the court. The interplay between orality and written medium is at the centre of the paper by Del Lungo Camiciotti and Bolton Holloway on the encounter of Margery Kempe and Julian of Norwich and shows how analyses carried out from different perspectives can reveal more on the meaning of such event than is textually attested. This section concludes with Petrina’s inspiring paper, which traces the influence of Petrarch’s works in courtly, aristocratic poetry in England, leading to the early-​sixteenth-​century flourishing of northern humanism, and thus, challenges the traditional idea of a divide between the late Middle Ages and the early modern period in English literature. The society and literature of Medieval England provide the background to the papers in the second part “Borrowing and the Lexicon”, which deals with loanwords in the Middle English lexicon. The three papers do not concentrate on the most frequently studied feature, that is, French influence, but largely focus on Old Norse borrowings: Lutz traces the attestations of the Old English word gærsum(a) ‘treasure’ of Norse origin, replaced with synonymous tresor but still, though more restrictedly, used in Middle English fictional and religious texts until the fifteenth century; whereas Molecki analyses the gradual lexical replacement of the verbs inherited from Old English (ge)limpan and befeallan) by the new formations from the Norse root happ-​, i.e., happe(n) and happen(en), in the late fourteenth and in the fifteenth century. The conclusive paper by Sylvester, Parkin and Ingham reports how changes in semantics and lexicon differ at the various levels of the lexical hierarchy, with a particular focus on technical language, in particular, on terms for Building, Domestic activities, Farming,

Preface

11

Food preparation, Manufacture, Trade, and Travel by water, through a quantitative analysis of the data collected for the Bilingual Thesaurus of Everyday Life in Medieval England. The five papers in the third section “Language at different levels” analyse different phenomena pertaining to different levels and functions of language structures. Mäkinen proposes a method of remapping linguistic items from different meta-​ discourse classes onto the Aristotelian rhetorical concepts, through a qualitative-​quantitative analysis of the Middle English translation of Henry Daniel’s Latin prologue to his English translation of Liber uricrisiarum. Stenbrenden examines back spellings (analogical spellings) in the history of English, applying the concept of “literal substitution sets”, and shows that the motivations for such spellings include not only phonemic merger, but also analogy, calligraphic changes and orthographic extensions, phonetic as well as phonological innovations, and that the phenomenon cannot be completely understood if scribal behaviour and clashes between scribal and authorial systems, both written and spoken, are not taken into account. From a more theoretical perspective, Stenroos reflects on how the implications of applying the dialect area and dialect continuum on Middle English language variation in writing, as different ways of inquiry, bring about different dialect maps, and suggests that an alternative approach to linguistic variation based on real historical contexts might reveal unexpected patterns. These patterns do not delineate a “pure dialect”, but socio-​linguistic facts, such as peoples’ geographical movements, and literary texts in a different light. The last two contributions go back to the phenomena of morphological loss and replacement during the Middle English period (1100–​1500), with Wårvik concentrating on the substitution of the Old English ambiguous adverb/​conjunctions þa and þonne for the adverb then and the conjunction when and the consequent redistribution of their syntactic and discourse-​pragmatic functions. Finally, Zimmermann investigates the disappearance of the negative particle ne in its function as a sentential negator and in negative concord structures, on the bases of to a quantitative analysis of data from three syntactically parsed Middle English corpora. It is my pleasure and duty to express my deep gratitude to the colleagues at the University of Florence as well as to the students, whose collaboration significantly contributed to the success of ICOME11. Moreover, I would like to thank all the colleagues involved in the peer review process, both at the conference abstract stage and all later stages, for their invaluable contribution. I would like to thank the University of Florence, the City Council of Florence and the Regione Toscana for their sponsorship, the Rector of the University of Florence for offering the venue.

newgenprepdf

12

Preface

I am grateful to the department of Education, Language, Interculture, Literature and Psychology (FORLILPSI) at the University of Florence for providing general encouragement for this book project and financially supporting its publishing through the project ICOME11. I would also like to thank the series editor, Professor Magdalena Bator, as well as Adam Gorlikowski and Łukasz Gałecki at Peter Lang, for their help, assistance and patience. Above all, I would like to thank the authors, both for their stimulating papers and for their collaboration during the various phases of the editing of this volume. Florence, March 2021 The Editor

Part I:  Textual Interlacing

Claudio Cataldi

A Re-​assessment of Poema Morale and its Influence on Penitence for Wasted Life Abstract: The aim of this study is to re-​assess the possible influence of Poema Morale on the slightly later lyric Penitence for Wasted Life. The intention is to consider both the content and metre of the two works. Previous scholarship has noted that Penitence for Wasted Life is thematically close to the early Middle English poem; as I shall show, this debt extends to metre as well. A wise old man’s reflection on the transience of worldly things, Poema Morale displays a fondness for proverbial sayings and vivid descriptions of heaven and hell –​all elements that must surely have appealed to the Early Middle English readership. This appeal is attested to not only by the nine manuscripts in which the poem is preserved, but also by several textual borrowings from Poema Morale in a number of thirteenth-​century lyrics, which were noted by previous scholarship. In this study, I shall suggest that, amongst these lyrics, Penitence for Wasted Life seems inspired by a specific section of Poema Morale, and that several previously unnoticed metrical correspondences between the two works actually indicate that the author of Penitence for Wasted Life possessed a first-​hand knowledge of the twelfth-​century poem. Keywords: Early Middle English poetry, religious poetry, Early Middle English literature, Poema Morale, metre

1. Introduction The twelfth-​century Poema Morale has been frequently regarded by scholars as a text that exerted a remarkable influence on medieval English religious literature.1 This supposed influence is twofold. On one level it involves themes (such as the transience of worldly wealth and the visions of heaven and hell) that are often developed in twelfth-​and thirteenth-​century poetry. On another level, it is reflected by the metre employed in Middle English religious poems and lyrics, which sometimes combine the septenary (that is, the metre of Poema Morale) with other metres. In this essay, I shall attempt to show that at least one lyric (whose modern title is Penitence for Wasted Life) can be considered to be inspired by Poema Morale beyond doubt, because its contents and prosody are modelled 1 For an overview, see Hill (1977); detailed references to scholarship on the poem are discussed below.

16

Claudio Cataldi

after the twelfth-​century poem. To do so, I shall first introduce Poema Morale and its manuscript tradition; I shall then re-​assess the various correspondences between the poem and other texts of the period. Finally, I shall discuss the textual and metrical analogies between Poema Morale and Penitence for Wasted Life.

2. Poema Morale: Manuscript Context Poema Morale is the title commonly given to a religious poem from the late twelfth century. Also known as A Moral Ode or The Conduct of Life, the poem is a reflection of a wise old man on his life, on the twin themes of sins and redemption, as well as on the transience of earthly things. This verse meditation is reinforced by a vivid description of the torments of hell and the joys of heaven. Poema Morale survives in nine manuscripts:2 1. Cambridge, Fitzwilliam Museum, MS McLean 123 (Nuneaton, West Midlands, s. xiii/​xiv). The text of Poema Morale in MS McLean 123 (fols 115r–​120r) is prefaced by two lines from another early Middle English poem, known as Sinners Beware!; along with names of English graphs on fol. 114v, this is the sole English material in a codex otherwise preserving French and Latin works.3 2. Cambridge, Trinity College, MS 335 (B.14.52) (East Midlands, late s. xii). The text of Poema Morale (fols 2r–​9v) precedes the collection of anonymous homilies known as Trinity Homilies.4 3. Durham, University Library, MS Cosin V.III.2 (England, early s. xiii). MS Cosin V.III.2 mostly includes Latin sermons, along with two lines from Poema Morale on fol. 127va/​5–​7:5 “suete bet swines brede ant of wilde dere. harde ye hus abiet hat haruore gift hiis swire” (grilled meat of swine and wild deer is sweet, but he who gives the neck for it pays it too dearly),6 corresponding to lines 145–​146 of the poem (Laing 1993: 53). 4. London, British Library, MS Egerton 613 (South-​West, early s. xiii). A religious miscellany, MS Egerton 613 includes two versions of the poem (fols

2 3 4 5

A list of the manuscripts of the poem is in Laing (1993: 162). On the manuscript, see Hill (2002); Laing (2013a). On the manuscript, see especially Laing –​McIntosh (1995); Hill (2003a). See Durham University Library, www.dur.ac.uk/​library/​asc/​theme/​medmss/​apviii2 [accessed 27 June 2020]. 6 Translations are my own.

A Re-assessment of Poema Morale

17

7r–​12v and 64r–​70v, respectively), which can be considered as two separate witnesses.7 5. London, British Library, MS Royal 7 C.iv (Christ Church, Canterbury, s. xi med/​s. xii/​xiii). The manuscript features late Old English continuous interlinear glosses to Defensor’s Liber scintillarum (Kato 2010–​2013). On fols 106v–​107v, originally blank, a later hand added some glosses and a two-​line quote in MS Royal 7 C. iv, partly erased today: “elde me is bestolen on er [...] Ne mæg ic geseo before me” (Ker 1957: 323), which corresponds to lines 17–​18 of Poema Morale: “Elde me is bistolen on ar ich hit iwiste | Ne mai ich isien bifore me for smeche ne for miste” (Old age has sneaked upon me before I was aware of it, and I cannot see before me because of smoke and mist) (Hall 1920: 31).8 6. London, Lambeth Palace Library, MS 487 (West Midlands, late s. xii/​early s. xiii). The codex of the early Middle English homiletic collection known as Lambeth Homilies; the text of Poema Morale is found at fols 59v–​65r (Sisam 1951). 7. Maidstone, Maidstone Museum, MS A.13 (Northampton, East Midlands, s. xiii). On fol. 93r of the manuscript is found the same couplet of Poema Morale as in MS Cosin v.iii.2, along with a shortened version of the Proverbs of Alfred.9 This couplet from Poema Morale also appears in two other folios of the manuscript, 46v and 253r (cf. Laing 1993: 120 and Laing 2013b). 8. Oxford, Bodleian Library, MS Digby 4 (Kent/​Surrey? early s. xiii). The text of Poema Morale (fols 97r–​110v) is the sole English material in a manuscript otherwise preserving Latin texts (see Laing 1993: 127). 9. Oxford, Jesus College, MS 29, part II (South-​West Midlands, s. xiii). A collection of mostly Middle English poems that, besides one of the two surviving copies of The Owl and the Nightingale, features the Proverbs of Alfred, Poema Morale (169r–​174v) and religious poems such as Passion of Our Lord, Sinners, Beware!, Doomsday, and Latemest Day, all of which will be discussed below (see Hill 1963 and Laing 2013c). The versions of Poema Morale recorded in these manuscripts vary in length, spanning from the nearly 400 lines in MS Trinity College 335 to the two lines

7 See Hill (1966); see also British Library, www.bl.uk/​catalogues/​illuminatedmanuscripts/​ record.asp?MSID=8935&CollID=28&NStart=613 [accessed 27 June 2020]. 8 For a discussion of these lines, see Hill (1977: 117). 9 See Laing (1993: 120), who also notes that the couplet is “close to the version in Lambeth 487”; see also Hill (2003b).

18

Claudio Cataldi

in MSS Cosin V.III.2, Royal 7 C iv, and Maidstone Museum A.13 –​which may be better described as quotes rather than witnesses of the text. The tradition of the two-​line quotes suggests that excerpts of the poem were used as proverbs of their own. The nine versions of Poema Morale also differ in variant readings and line order. These differences presumably account for the relative scarceness of critical editions and secondary literature on the poem. Lewin (1881) attempted at a critical edition based on the six manuscripts known at the time; Marcus (1934) published a critical text based on seven versions (including the two different witnesses in MS Egerton 613). However, Hill (1977: 100–​101) considers Marcus’ edition to be unsatisfactory, calling it a “personal reconstruction of a text whose place and date of origin is not proved”. The main manuscript versions have also been printed separately. Morris included the Jesus College 29 version in his Old English Miscellany (Morris 1872: 58–​71) and the versions from MSS Trinity College 335 and Lambeth Palace 487 in his editions of the Trinity Homilies (Morris 1873: 220–​232) and the Lambeth Homilies (Morris 1867: 158–​183), respectively. Hall (1920: 30–​53, 312–​354) prints the Trinity and Lambeth versions and offers detailed notes on the poem. Given that it represents a full copy of the poem, and one close to the archetype –​as discussed by Hill (1977: 114–​115) in what is perhaps the most complete study of the poem to date –​I shall use the Trinity College 335 version of the poem as a base text for my discussion.10

3. Poema Morale and the Twelfth-​and Thirteenth-​Century Religious Poetry The fact that versions of (or quotes from) Poema Morale are preserved in nine manuscripts is an obvious measure of its popularity amongst Middle English readers. This is hardly surprising, because the poem exhibits a combination of elements that must surely have appealed to a late twelfth-​century audience: exhortations to the perfect Christian life, a vivid vision of Hell, and a concluding overview of the joys of eternal life, along with a fondness for words of wisdom and proverbial sayings, all of which are frequently found in other religious poems of the same period (Hill 1977: 116–​126). The popularity of Poema Morale seems also proven by a number of analogues to the poem. Previous scholarship has noted the resemblance between Poema Morale and Le romaunz de temtacioun de secle, an early twelfth-​century verse sermon (attributed to an Anglo-​Norman

10 As printed by Hall (1920: 30–​53).

A Re-assessment of Poema Morale

19

author, Guischart de Bealiau), which is preserved in four manuscripts of English provenance.11 The Romaunz and Poema Morale share a remarkable number of parallels, which have been included in a study by Gabrielson (1912) and in the notes to the edition of the early Middle English poem by Hall (1920: 329–​354). That some sort of textual relationship between the two works exists seems to this author to be beyond reasonable doubt; the nature of this relationship is, however, far from certain. Gabrielson (1912: 311–​312) argues that certain ideas point towards a shared background from the two authors, rather than direct influence; however, he also suggests that the poet of the Romaunz may have recalled verses from Poema Morale from memory.12 Scholars such as Lewin (1881), Patterson (1911), Brown (1932), and Marcus (1934) have identified several occurrences of verses apparently drawn from Poema Morale amongst the corpus of early Middle English poetry. In particular, these occurrences are found in the following texts: 1. Doomsday, which survives in four manuscripts: Cambridge, Trinity College, MS B.14.39; London, British Library, MS Cotton Caligula A.ix; Oxford, Bodleian Library, MS Digby 86; and Jesus College 29;13 2. On God Ureisun of Ure Lefdi, a lai, related to the Wooing Group, preserved in Cotton Nero A.XIV; printed by Morris (1867: 190–​199), Hall (1920: 132–​ 137), and Brown (1931: 3–​8); first line: Cristes milde moder seynte marie; 3. Latemest Day, preserved in the same manuscripts as Doomsday and, according to Brown (1932: 187), written by the same author;14 4. The Passion of Our Lord, a long poem printed by Morris (1872: 37–​57) from MS Jesus College 29, first line: Ihereþ nv one lutele tale; 5. Penitence for Wasted Life, a lyric preserved in London, BL, MS Additional 27909, first line: Leuedi sainte marie, moder and meide; modern title given by Brown (1932: 1–​2);15 1 1 See Gabrielson (1909); Brun (2005–​2019a); Hill (1977: 123–​126). 12 Hill (1977: 123–​126) seems to agree with Gabrielson on the relationship between the Romaunz and Poema Morale, although she challenges the attribution of the former to Guischart de Beauliu. The web of textual relationship is actually complex: the Romaunz also recalls previous Old English religious literature, especially Ælfric –​echoes perhaps mediated by Middle English versions of his homilies (see Gabrielson 1912: 310–​311); both the Romaunz and Poema Morale parallel Anglo-​Norman works such as the verse sermon Grant mal fist Adam: see Gabrielson (1912: 310); Hall (1920: 314); edition in Suchier (1879). See also Brun (2005–​2020b). 13 Printed in Brown (1932: 42–​46). 14 Edition in Brown (1932: 46–​54). 15 The lyric is titled A Prayer to our Lady by Morris (1872: 192–​193).

20

Claudio Cataldi

6. Sinners, Beware!, also known as Sayings of St Bede, an early tail-​rhymed penitential poem that survives in two manuscripts, MS Jesus College 29 (where it follows Poema Morale) and MS Digby 86 (one of the manuscripts of Doomsday and Latemest Day);16 7. This World’s Bliss Will Not Last, from London, British Library, MS Arundel 248; another lyric printed and titled by Brown (1932: 78–​82), first line: Worldes blis ne last no throwe. The main problem with assuming a direct influence of Poema Morale on these works (as well as on other lyrics) is that the poet of Poema Morale makes use of a variety of stock themes.17 Therefore, it is not easy to decide whether textual similarities indicate direct influence or rather point towards a similar treatment of the same theme. For example, this is the case of Poema Morale, v. 6: “Þan ibiðenche me þar on wel sore ime adrade” (when I think about it I become very afraid), which echoes the opening line of Doomsday –​“wenne hi þenche on domes-​dai ful sore I me adrede” (when I think about Doomsday I become very afraid) –​(cf. Brown 1932: 188 and Marcus 1934: 6). The topos of ineffability developed in Poema Morale, vv. 289–​290: “Nemai non herte hit þenche ne tunge hit ne mai telle | hwu muchele pine ne hwu fele senden in helle” (no heart can think nor tongue can tell how much pain and how many torments are in hell), recurs in Latemest Day, vv. 29–​30 (cf. Brown 1932: 191 and Marcus 1934: 9) and Sinners, Beware!, vv. 38–​41 (cf. Marcus 1934: 9). These latter two poems deserve particular attention. Latemest Day is a poem in mono-​rhyme quatrains that can be divided in two parts: the first part is based on the theme of the “Address of the Soul to the Body” –​a staple of Old and early Middle English religious literature –​and borrows motifs and specific lines from the Soul’s Address to the Body in MS Worcester Cathedral F.174; the second part of the poem is a description of hell that seems to have been inspired by the analogous passage in Poema Morale, vv. 220–​290.18 Similarly, the poet of Sinners, Beware! may well have had at hand a variety of penitential lyrics to draw inspiration from (as shown by the passage related to the theme of the “Soul’s Address to the Body” at vv. 331–​ 336), so that resemblances between Poema Morale and Sinners, Beware! probably

16 The version in MS Jesus College 29, incipit Þeos holy gostes myhte, is printed by Morris (1872: 72–​83). The MS Digby 86 version, first line: Holi gost, þi miȝtte, is in Furnivall (1901: 765–​776). 17 On the use of stock themes in Poema Morale, see Hill (1977: 119–​121). 18 On the parallels between Latemest Day and Poema Morale, see Hall (1920: 345); Brown (1932: 190–​191); Marcus (1934: 8–​9). See also Cataldi (2018: 138–​149).

A Re-assessment of Poema Morale

21

result from analogous treatments of the same stock themes.19 Parallels between Poema Morale on the one hand and Doomsday20 and The Passion of Our Lord21 on the other are not conclusive enough to prove a direct textual relationship. With regard to This World’s Bliss Will Not Last, Brown (1932: 201) has noted that “one hardly expects a direct literary source for a poem on such an oft-​handled theme as the transitoriness of worldly joy, but it is interesting to note a few lines which seem to be distinct echoes of the twelfth-​century Poema Morale” (cf. also Marcus 1934: 7). I find two of the parallels noted by Brown particularly convincing: This World’s Bliss, vv. 51–​52, “Scal no gud ben unforiolden, | ne no qued ne wrth unbout” (No good will be uncompensated nor wickedness will be unrewarded) echoes Poema Morale, v. 59, “Ne sal þar non euel ben unboht ne god unforȝolden” (No evil will be unrewarded nor good uncompensated); even more striking is the correspondence between Poema Morale, vv. 319–​320, “We wilnieð after wereldes wele þe longe ne mai ilaste | 7 legeð mast al ure swinc on þing unstedefaste” (we seek for the world’s wealth, that may not last long, and devote most of our work to transitory things), and This World’s Bliss, vv. 31–​34: Man, wi sestu þout and herte o werldes blis þat nout ne last? Wi þolstu þat þe softe ismerte for þing þat is unstedefast? (Brown 1932: 79) ‘O man, why do you set your thought and heart on worldly bliss that won’t last? Why do you endure to suffer so often for a thing that is transitory?’

The treatment of the theme of the transience of worldly goods –​framed by the rhyme last/unstedefast –​suggests that the poet of This World’s Bliss may have had Poema Morale in a corner of his mind; by contrast, some differences in wording (“wereldes wele” /​“werldes blis”; the “softe ismerte” passage) seem to exclude direct quotation but rather indicate –​as in the case of the Romaunz –​ that these verses are the product of a poet who recalled verses of Poema Morale from memory.

19 To such an extent that, in MS McLean 123, Poema Morale is introduced by an excerpt from Sinners, Beware! (see above). On the parallels between Sinners, Beware! and Poema Morale, see Hall (1920: 331); Brown (1932: 188); Marcus (1934: 7–​9). 20 On the parallels betweenDoomsday and Poema Morale, see Brown (1932: 188); Marcus (1934: 6). 21 See Hall (1920: 343) and Marcus (1934: 7–​8, 10) for some parallels between the poems.

22

Claudio Cataldi

4. Penitence for Wasted Life and Poema Morale As I now aim to show, Penitence for Wasted Life is indeed a very promising candidate for a work directly informed by Poema Morale. Penitence for Wasted Life is a thirteenth-​century prayer to the Virgin that recalls other Marian invocations of the period, for example the aforementioned On God Ureisun of Ure Lefdi.22. Surviving in a single manuscript (London, British Library, MS Additional 27909, fol. 2r) and written out as prose (Laing 1993: 63), it is a lyric in mono-​rhyme quatrains. So far, Penitence for Wasted Life has not received a great deal of scholarly attention; however, its commentators have almost universally noted its resemblance to Poema Morale.23 I now wish to focus on the fact that Penitence for Wasted Life seems to borrow from a specific portion of Poema Morale –​ that is, the opening section of the poem; moreover, the metre of Penitence for Wasted Life often imitates that of Poema Morale. Defined as “a personal lament of an old man looking back on a wasted and sinful life” by Duncan (1992: 109), Penitence for Wasted Life is to all intents and purposes a variation on the same meditative themes of Poema Morale, and almost an expansion of its first twenty-​one verses. I quote the relevant parallels in Table 1, grouping together all references to previous scholarship.24

22 On the Ureisun and the cult of Virgin Mary see Hall (1920: 537). It should be recalled that Marcus (1934: 9–​10), notes two textual echoes of Poema Morale in On God Ureisun of Ure Lefdi: the image of the “day without night” (Poema Morale, l. 370; Ureisun, l. 59) and the vision of divine beauty (Poema Morale, l. 392; Ureisun, l. 30). However, I think that these two parallels are not conclusive enough to prove a direct textual relationship; these images may derive from a similar background and are attested to elsewhere (see, for example, Hall 1920: 538). 23 In studies as early as Ten Brink (1883: 206–​207). Cf. Patterson (1911: 166); Hall (1920: 330); Brown (1932: 166); Marcus (1934: 6–​7). Booker (1912: 34) notes a “French influence” on the lyric. 24 The text of Penitence for Wasted Life is quoted from Brown (1932: 1–​2); ƿ is changed into w.

23

A Re-assessment of Poema Morale Table 1:  Correspondences between Poema Morale and Penitence for Wasted Life 1

2

3

4

5

6

7

Poema Morale 5–​6: Vnnet lif ic habbe ilad 7 ȝiet me þincheð ilade | Þan ibiðenche me þar on wel sore ime adrade ‘I have led, and I think I still lead, a life in vain; when I think about it, I become very afraid’ 5–​6: Vnnet lif ic habbe ilad 7 ȝiet me þincheð ilade | Þan ibiðenche me þar on wel sore ime adrade ‘I have led, and I think I still lead, a life in vain; when I think about it, I become very afraid’ 10: 7 fele ȝeunge dade idon þe me ofðinkeð nuðe ‘And I have committed many immature deeds which I now regret’ 11: Alto lome ich habbe igult a werke 7 a worde ‘I’ve done wrong too frequently with acts and words’ 11: Alto lome ich habbe igult a werke 7 a worde ‘I’ve done wrong too frequently with acts and words’ 12: Alto muchel ic habbe ispend to litel ileid on horde ‘I’ve spent too much and I’ve saved too little in the hoard’

Penitence for Wasted Life 3–​4: Vnnut lif to longe ich lede; | hwanne ich me biþenche wel sore ich me a-​drede ‘I have led a life in vain for too long; when I think about it, I become very afraid’

17: Elde me is bistolen on ar ich hit iwiste ‘Old age has sneaked upon me before I was aware of it’

9: Slep me hað mi lif forstole richt half oðer more ‘Sleep has stolen right half of my life, or more, from me’

Noted by Lewin (1881: 44–​45) Patterson (1911: 166) Hall (1920: 330) Brown (1932: 166) Marcus (1934: 6) 44: Hwan ich hier-​of rekeni Lewin schal, wel sore me mei drede (1881: 44–​45) ‘When I will have to render an Marcus (1934: 6) account [of my conduct] here, I will become very afraid’

31: & wel feole sunne ido þe me ofþincheð nuðe ‘And many sins which I now regret’

Lewin (1881: 44–​45) Marcus (1934: 6)

21: Ifurn ich habbe isunehed mid worke & mid worde ‘I’ve committed sins for so long with acts and words’

Patterson (1911: 166) Hall (1920: 330) Marcus (1934: 6)

29 Ifurn ich habbe isuneȝet mit wurken and midd muðe ‘I’ve committed sins for so long with acts and mouth’

Lewin (1881: 44–​45) Patterson (1911: 166)

24: Muchel ich habbe ispened, to lite ich habbe an horde ‘Much I have spent, too little I have in the hoard’

Patterson (1911: 166) Brown (1932: 166) Marcus (1934: 7) Patterson (1911: 166)

(continued on next page)

24

Claudio Cataldi

Table 1: Continued 8

9

Poema Morale 17: Elde me is bistolen on ar ich hit iwiste ‘Old age has sneaked upon me before I was aware of it’ 21: Þe wel ne deð þe hwile he mai wel ofte hit sal him rewen ‘Those who don’t act well while they can, shall regret it very often’

Penitence for Wasted Life 17: Slep me hað mi lif forstole er ich me bisehe ‘Sleep has stolen my life from me before I paid attention to it’ 10: Awai! to late ich was iwar, nu hit me reoweð sore ‘Alas, I have become aware of it too late, and now I regret it sorely’

Noted by Brown (1932: 166) Patterson (1911: 166)

Quotes 1, 3–​4, 6 agree in both content and wording; 2, 5, 7–​8 feature some differences in wording which are, in turn, reflected in their content (for example elde/​slep in 6–​7); 9 has similar content but different wording. Notably, all the lines of Penitence for Wasted Life that parallel Poema Morale are analogous to a specific portion of the poem, that is, the first 21 lines. In other words, it seems that the poet of Penitence for Wasted Life borrowed a batch of lines from the opening section of Poema Morale. As the two-​line excerpts mentioned above show, excerpts from Poema Morale are attested to independently as maxims (cf. Hill 1977: 377); therefore, it is likely that portions of the text circulated separately and influenced some religious poems of the period.

5. Metre Further evidence of a direct influence of Poema Morale on Penitence for Wasted Life comes from metre. To adequately discuss this issue, it is worth reminding that Poema Morale is usually considered to be the first English poem composed in septenaries (cf. Macdonald Alden 1903: 260; Schipper 1910: 193; Hall 1920: 327–​329). The septenary is a metre derived from Latin poetry and based on seven-​feet syllabic lines divided in four-​stress and three-​stress hemistichs; the rhythm is usually trochaic, but quite often changed into iambic by means of the addition of an extra syllable at the beginning of each hemistich (Hall 1920: 327). According to Hall (1920: 327–​329, 486), Poema Morale reflects the influence of native prosody and the use of all possible ‘licences’ to fit the standard succession of stressed and unstressed syllables required by its Latin model; such as elisions, syncopations, and hiatuses of vowels, as well as the occasional addition of unstressed syllables. The metre of the poem has also been recently discussed by Fulk (2002) and Minkova (2016), whose studies identify prominent prosodic

A Re-assessment of Poema Morale

25

elements of Poema Morale by the following features: the frequent presence of a strong caesura between the hemistichs –​but with the possibility of an unstressed syllable at the beginning of the second hemistich –​; elision of final unstressed -​e before vocalic or weak-​h initials; penultimate heavy syllable; syncopation of weak syllables in trisyllabic words; and a rhythm that is predominantly iambic.25 All these features result in a variety of verse possibilities. By comparison, the use of septenaries in the nearly-​contemporary Orrmulum shows a systematic succession of stressed and unstressed feet in fifteen-​syllable lines; the Orrmulum also lack end-​rhymes, while Poema Morale rhymes in couplets (see Hall 1920: 327 and Hill 1977: 115–​116; Pearsall 2018: 33–35). The septenary went on to be employed throughout early and later Middle English religious literature: it is the metre of the South English Legendary; of a section of Cursor Mundi; and of the Life of St Margaret (see Hill 1977: 115). In his English Verse, Macdonald Alden (1903: 253–​254) noted that, in early Middle English poetry, septenaries were often used interchangeably with alexandrines, quoting Doomsday, Latemest Day, and the Passion of Our Lord as examples of this tendency; he also argued that “the alexandrine was easily confused by the Middle English writers, not only with the septenary, but with the native “long line”, and it is often difficult to say just what rhythm was in the writer’s mind”. In fact, the authors of poems such as Doomsday, Latemest Day, and the Passion of Our Lord probably aimed at reproducing a rhythm rather than a metrical form. This situation is well exemplified by On God Ureisun of Ure Lefdi, a text considered by Hall (1920: 536) as employing a mixture of septenaries, alexandrines, and alliterative long lines; Hall convincingly talks of “alliterative long line in the last stage of its dissolution”. More recently, Conlee (1991: 11) has called the septenary “a staple of thirteenth-​and fourteenth-​century homiletic writing, which often includes variants with six stresses and occasionally shows influence of the four-​stress alliterative long line” (see also Macdonald Alden 1903: 260–​261). For present purposes, it is worth asking ourselves whether Poema Morale may have played a role in the development of the rhythmical mixture of alexandrines, septenaries, and alliterative long lines found in early Middle English religious poems such as Latemest Day, the Passion of Our Lord and On God Ureisun of Ure Lefdi. According to Hill (1977: 115) the poem “had important effects on English poetry. It demonstrated that the penitential mood […] could be expressed in

25 Fulk (2002: 346–​353) also suggests that some lines of the poem may show the influence of resolution, a feature which, he argues, may derive from Old English prosody; this hypothesis is further discussed by Minkova (2016: 122–​143).

26

Claudio Cataldi

a new adaptable verse form”. The poet of Poema Morale skilfully uses the sep­ tenary throughout its work, without “contaminations” from other verse forms; but other poets may well have done so.26 Obviously, the fact that the poet of Poema Morale used the septenary for religious poetry does not necessarily imply that it directly influenced other poets, unless we find that textual parallels correspond to metrical similarities. This is exactly the case of the correspondences between Poema Morale and Penitence for Wasted Life. To my knowledge, it has passed unnoticed that Penitence for Wasted Life features some lines that are clearly modelled after the septenary. I quote the relevant lines in Table 2.27 Table 2:  Septenaries in Penitence for Wasted Life Verse no. 4 9 10 13 15 16 17 19 20 21 29 30 31 32 36 39 41 43 44

Line hwánne ich | mé bi|þén|che wél || sóre ich| mé a-​|dréde slép me | háð mi | líf for|stóle || rícht half | óðer | móre awái! | to lá|te ich wás | iwár, || nu hít | me réoweð | sóre ál to | lónge | slépð þe | mánn || þat néure | néle a|wákie to dón|de súnn|e awéi | fram hím || and féle al|mésse | mákie ȝif hím | ne schál | hwanne hé | forð-​wánt || his bréi-​|gúrdel | quákie slép me | háð mi | líf for|stóle || ér ich | mé bi|séhe mi brú|ne hér | is hwít | bicúme || ich nót | for hwúcche | léihe ánd mi | tóhte | rúde i|túrned || al ín-​|to óðre | déhe ifúrn | ich hább|e isúne|héd || mid wórke | ánd mid | wórde ifúrn | ich hább|e isúne|ȝét || mit wúrken | ánd midd | múðe ánd mid | álle | míne | líme || síððe ich | súnehi | cúðe ánd wel | féole | súnne i|dó || þe mé of|þíncheð | núðe and swó | me hádd|e ifúrn | idó || ȝif hít | me críst i-​|ȝúðe lúue to | góde | ánd te | mánn || ic bídde | þat tú | me sénde hwas fléch | and blód | ihál|ȝed ís || of bréd | of wáter | of wíne ínne | méte and | ínne | drínke || ic hábbe i|béo ouer|déde hwánne ich | ihúrde of | góde | spéke || ne hédd | ich hwát | me séde hwán ich | hiér-​of | rékeni | schál || wel sóre | mé mei | dréde.

26 Textual corruption further complicates the matter. Hall (1920: 536): “a scribe dealing with an older text was generally little concerned about the form and much about the matter and the transcription of its language into his own dialect and idiom”. 27 Text quoted from Brown (1932: 1–​2); the abbreviation & is expanded into and for metrical reasons; ƿ is changed into w.

A Re-assessment of Poema Morale

27

Septenaries are used throughout four whole quatrains (13–​16, 17–​20, 29–​32, 41–​44), as well as in a number of single lines; lines are made up of four-​syllable + three-​syllable hemistichs. There is an alternance between strong (masculine) caesuras, as in ll. 4, 10, 13 15, 16, and first hemistichs with feminine endings (that is, an unstressed vowel after the fourth stress), as in ll. 9, 17, 19, 30; the second hemistich frequently begins with an unstressed vowel (ll. 10, 13, 19); regardless of rhythm, the second hemistich always has a feminine ending. There is frequent elision of unstressed vowels before words with vocalic initials (ll. 4 “hwanne ich”, 10 “late ich”, 20 “rude iturned”, 30 “siððe ich”, 31 “sunne ido”, 41 “habbe ibeo”), but this never occurs with me + vocalic initials (for example l. 4 “me adrede”). There is syncopation of unstressed e (ll. 30 sunehi, 39 water); –​all features that occur in Poema Morale (cf. Hall 1920: 328–​329 and Fulk 2002: 344–​345). Notably, some of the verses quoted above (ll. 4, 9, 10, 17, 21, 29, 31, 44) are also those corresponding to the first section of Poema Morale (numbers 1–​5, 7–​9 in the table of correspondences). This means that a group of lines share both content and metre. Furthermore, the possibility that other lines were originally septenaries and were corrupted in an intermediate stage of tradition cannot be ruled out.28 For example, line 24 “Muchel ich habbe ispened, to lite ich habbe an horde” (number 6 in the table of correspondences above) features 3 stresses on each half-​line; this line might have been originally opened by “al to”, as in Poema Morale, line 12, thus being a septenary. However, even if we confine ourselves within the boundaries of the existing text, the textual parallels and the presence of septenaries indicate a first-​hand knowledge of Poema Morale, or at least of excerpts from the poem, by the author of Penitence for Wasted Life. This partially confirms the hypothesis of the influence of Poema Morale on the use of septenary as a metre for early Middle English religious poems.29

6. Conclusions Cases of proven textual dependence among texts do not abound in early English literature; in this context, we can assume that Poema Morale directly inspired Penitence for Wasted Life both in content and metre. My hypothesis is that a batch of lines from the opening section of the twelfth-​century poem were reworked into the thirteenth-​century lyric, which was also modelled after the 28 Duncan (1992: 109–​111), has already suggested that the surviving text of Penitence for Wasted Life shows signs of textual corruption, such as the inversion of two quatrains. 29 Penitence for Wasted Life also shows surviving traces of alliteration; see line 1: Leuedi sainte marie, | moder and maide; line 8: wielde Godd an heuene to hwuechere wunne!

28

Claudio Cataldi

metre of its source. This suggests that the poet of Penitence for Wasted Life had direct access to a copy of at least the opening section of Poema Morale. In a wider context, the influence of Poema Morale deserves re-​assessment. As discussed above, excerpts from the text were used as maxims; the study of Penitence for Wasted Life demonstrates that specific sections of the poem were reworked by later authors; furthermore, it also shows that the hypothesis of an influence of the poem on the development of septenary as a metre for religious poetry is grounded; echoes of Poema Morale in other texts (for example, Le romaunz de temtacioun de secle and This World’s Bliss Will Not Last) also suggest that the poem was presumably quoted from memory by other authors. Therefore, the place of Poema Morale in early English literature is perhaps even more central than previously thought: its mixture of wisdom and vision, assembled in a new and fashionable metre, enjoyed an enduring popularity throughout early Middle English literature.30

References Beadle, Richard –​A. J. Piper (eds.). New Science out of Old Books: Studies in Manuscripts and Early Printed Books in Honour of A. I. Doyle. Aldershot: Scholar Press. Booker, John M. 1912. A Middle English Bibliography. Heidelberg: C. Winter. Brown, Carleton (ed.). 1932. English Lyrics of the Thirteenth Century. Oxford: Clarendon Press. British Library. “Egerton 613”. In: Catalogue of Illuminated Manuscripts. http://​ www.bl.uk/​catalogues/​illuminatedmanuscripts/​record.asp?MSID=8935&Col lID=28&NStart=613 [accessed 27 June 2020]. Brun, Laurent. 2005–​2020a. “Guischart de Beauliu”. In: Brun, Laurent et al. (eds.). Archives de littérature du Moyen Âge (ARLIMA). http://​www.arlima. net/​eh/​guischart_​de_​beauliu.html [accessed 27 June 2020]. Brun, Laurent. 2005–​2020b. “Grant mal fist Adam”. In: Brun, Laurent et al. (eds.). Archives de littérature du Moyen Âge (ARLIMA). http://​www.arlima.net/​eh/​ grant_​mal_​fist_​adam.html [accessed 27 June 2020]. Cataldi, Claudio. 2018. “A Literary History of the ‘Soul and Body’ Theme in medieval England”. unpublished PhD dissertation, University of Bristol. 30 I am most grateful to Professor Ad Putter (University of Bristol) and Professor Patrizia Lendinara (Università degli Studi di Palermo), as well as to the anonymous readers who commented on earlier drafts of this paper, for suggesting improvements. I am solely responsible for any errors or infelicities.

A Re-assessment of Poema Morale

29

Conlee, John W. (ed.). 1991. Middle English Debate Poetry. A Critical Anthology. East Lansing, MI: Colleagues Press. Duncan, Thomas G. 1992. “Textual Notes on Two Early Middle English Lyrics”. Neuphilologische Mitteilungen 93: 109–​120. Durham University Library. “DUL MS Cosin V.III.2”. In: Medieval Manuscripts in the University Library. http://​www.dur.ac.uk/​library/​asc/​theme/​medmss/​ apviii2 [accessed 27 June 2020]. Fulk, Robert D. 2002. “Early Middle English Evidence for Old English Meter: Resolution in Poema Morale”. Journal of Germanic Linguistics 14: 331–​355. Furnivall, Frederick J. (ed.). 1901. The Minor Poems of the Vernon MS, Part 2. EETS OS 117. London: Kegan Paul, Trench, Trübner. Gabrielson, Arvid (ed.). 1909. Le sermon de Guischart de Beauliu. Édition critique de tous les manuscrits connus avec introduction. Uppsala: Akademiska Bokhandeln; Leipzig: Harrassowitz. Gabrielson, Arvid. 1912. “Guischart de Beauliu’s Debt to religious Learning and Literature”. Archiv für das Studium der neueren Sprachen und Literaturen 128: 309–​328. Hall, Joseph (ed.). 1920. Selections from early Middle English 1130–​1250. 2 vols. Oxford: Clarendon Press. Hill, Betty. 1963. “The History of Jesus College, Oxford MS 29”. Medium Ævum 32: 203–​213. Hill, Betty. 1966. “Notes on the Egerton e-​Text of the Poema Morale”. Neophilologus 50: 352–​359. Hill, Betty. 1977. “The Twelfth-​Century Conduct of Life, Formerly the Poema Morale or A Moral Ode”. Leeds Studies in English 9: 97–​144. Hill, Betty. 2002. “A Manuscript from Nuneaton: Cambridge, Fitzwilliam Museum MS McLean 123”. Transactions of the Cambridge Bibliographical Society 12: 191–​205. Hill, Betty. 2003a. “Cambridge, Trinity College MS B.14.52”. Transactions of the Cambridge Bibliographical Society 12: 393–​402. Hill, Betty. 2003b. “A Couplet from the Conduct of Life in Maidstone MS A 13”. Notes and Queries 50: 377. Kato, Takako. 2010–​2013. “London, British Library, Royal 7 C. iv”. In: Da Rold, Orietta –​Takako Kato –​Mary Swan –​Elaine Treharne (eds.). The Production and Use of English Manuscripts 1060 to 1220. Leicester: University of Leicester. http://​w ww.le.ac.uk/​english/​em1060to1220/​mss/​EM.BL.Roya.7.C.iv.htm [accessed 27 June 2020].

30

Claudio Cataldi

Ker, Neil R. 1957. Catalogue of Manuscripts Containing Anglo-​ Saxon. Oxford: Clarendon Press. Laing, Margaret. 1993. Catalogue of Sources for A Linguistic Atlas of Early Medieval English. Cambridge: Brewer. Laing, Margaret –​Angus McIntosh. 1995. “Cambridge, Trinity College MS 335: Its Texts and Their Transmission”. In: Beadle, Richard –​A. J. Piper (eds.): 14–​52. Laing, Margaret. 2013a. “Cambridge, Fitzwilliam Museum, McClean 123”. In: Laing, Margaret (ed.). http://​archive.ling.ed.ac.uk/​ihd/​laeme1_​scripts/​ search_​cross_​ref.php?fieldVal=Cambridge,%20Fitzwilliam%20Museum,%20 McClean%20123 [accessed 27 June 2020]. Laing, Margaret. 2013b. “Maidstone Museum A.13, entry 1”. In: Laing, Margaret (ed.). http://​archive.ling.ed.ac.uk/​ihd/​laeme2_​scripts/​search_​ cross_​ref.php?fieldVal=Maidstone%20Museum%20A.13,%20entry%201 [accessed 27 June 2020]. Laing, Margaret. 2013c. “Oxford, Jesus College 29, part II”. In: Laing, Margaret (ed.). http://​archive.ling.ed.ac.uk/​ihd/​laeme1_​scripts/​search_​ cross_​ref.php?fieldVal=Oxford,%20Jesus%20College%2029,%20part%20 II. [accessed 27 June 2020]. Laing, Margaret (ed.). 2013. A Linguistic Atlas of Early Middle English, 1150–​ 1325 (LAEME). Edinburgh: University of Edinburgh. (Version 3.2). http://​ www.lel.ed.ac.uk/​ihd/​laeme2/​laeme2.html [accessed 27 June 2020]. Lewin, Hermann (ed.). 1881. Das mittelenglische Poema Morale. Halle: Max Niemeyer. Macdonald Alden, Raymond (ed.). 1903. English Verse. Specimens Illustrating Its Principles and History. New York: Henry Holt and Company. Marcus, Hans (ed.). 1934. Das Frühmittelenglische Poema Morale. (Palaestra, 194). Leipzig: Mayer und Müller. Minkova, Donka. 2016. “Prosody-​Meter Correspondences in Old English and Poema Morale”. In: Neidorf, Leonard –​Rafael J. Pascual –​Thomas A. Shippey (eds.): 122–​143. Morris, Richard (ed.). 1867. Old English Homilies and Homiletic Treatises. (EETS OS 29 and 34). London: Trübner. Reprinted 1998. Morris, Richard (ed.). 1872. An Old English Miscellany: A Bestiary, Kentish Sermons, Proverbs of Alfred and Religious Poems of the 13th Century. (EETS OS 49). London: Trübner. Morris, Richard (ed.). 1873. Old English Homilies of the Twelfth Century. From the Unique Ms. B. 14. 52 in the Library of Trinity College, Cambridge. (EETS OS 53). London: Trübner.

A Re-assessment of Poema Morale

31

Neidorf, Leonard –​Rafael J. Pascual –​Thomas A. Shippey (eds.). 2016. Essays in Old English Philology: Studies in Honour of R. D. Fulk. Cambridge: D.S. Brewer. Patterson, Frank A. 1911. The Middle English Penitential Lyric: A Study and Collection of Early Religious Verse. New York: Columbia University Press. Pearsall, Derek. 2018. “The Metre of the Tale of Gamelyn”. In: Putter, Ad –​Judith A. Jefferson (eds.): 33–​49. Putter, Ad –​Judith A. Jefferson (eds.). 2018. The Transmission of Medieval Romance. Metres, Manuscripts and Early Prints. (Studies in Medieval Romance). Cambridge: D.S. Brewer. Schipper, Jacob. 1910. A History of English Versification. Oxford: Clarendon Press. Sisam, Celia. 1951. “The Scribal Tradition of the Lambeth Homilies”. The Review of English Studies 2: 105–​113. Suchier, Hermann (ed.). 1879. Reimpredigt. (Bibliotheca Normannica: Denkmäler Normannischer Literatur und Sprache 1). Halle: Max Niemeyer. Ten Brink, Bernhard. 1883. Early English Literature (to Wiclif). Translated from the German by Horace M. Kennedy. New York: Henry Holt and Company.

Gabriella Del Lungo Camiciotti –​Julia Bolton Holloway

The Soul a City: Margery meets Julian Abstract: This fittingly partnered essay discusses the meeting between Margery Kempe and Julian of Norwich, which took place around 1413. We knew froHm the Book of Margery Kempe that she was seeking advice from the Anchoress Julian to validate her form of spirituality. We therefore presented a double perspective on the meeting of the two women: the first discursive, focusing on orality and voices that we hear in the Book, demonstrated in our own oral reading that repeated their dialogue, while presenting the reasons for that meeting; the second, philological and linguistic, studying its unique extant manuscript deriving from the Mount Grace Charterhouse in Yorkshire copying out a lost original in the East Anglian dialect. Their reported dialogue centres on the Discernment of Spirits that swirled about Margery’s model, Birgitta of Sweden. Study of the actual manuscript, rather than its modern editions, can give a clearer sense of the psychological consolation received by Margery from Julian in this scribally documented meeting between two medieval women conversing with each other orally. Keywords: Margery Kempe, Julian of Norwich, spiritual encounter, female spiritual network, orality, scribalism

1. Introduction Margery Kempe visited Julian of Norwich around 1413 and later reported their conversations, thus providing for us an Oral Text (Allen –​Meech 1940 [1982], Berry 2000) to take its place alongside Julian’s surviving Westminster Text, Long Text and Short Text of her Showing of Love (Reynolds –​Holloway 2001). Margery’s Book allows us to go back to fifteenth-​century East Anglia with, as it were, a microphone and recording equipment. Julian functioned in her community much like a psychiatrist, healing souls; Julian helps heal Margery’s soul, perhaps, too, by suggesting the therapy of the Jerusalem pilgrimage and the writing of the vast book of her travels, The Book of Margery Kempe, these two women’s vernacular books imitating St Birgitta of Sweden’s Latin Revelationes (Birgitta 1997-​). We therefore presented orally a double perspective on the meeting of the two women: the first, discursive, focusing on orality, the voices of both women and the men whom we hear in the Book; the second, philological and linguistic, the text we read from taken diplomatically from the unique Butler-​Bowden manuscript, written in secretary hand with ampersands and rubricated letters,

34

Gabriella Del Lungo Camiciotti & Julia Bolton Holloway

now British Library Additional 61823, rather than from an edited modern printed text. We chose to focus on these two English women who actually met in Norwich: Margery Kempe and Julian of Norwich. The books they wrote present different ways of living their religious calling and illustrate two aspects of late medieval piety: that of Julian, who was a recluse dwelling in an anchorhold attached to Saint Julian’s Church in Norwich, who had few but profound visions she then studied, and who has been defined as a theologian (Pezzini 2015: 61–​ 62). The other, instead, of Margery, a married woman, with fourteen children, who became a much travelled pilgrim and who continually practiced imaged “Sacred Conversations”, reaching back to St Paula, observed by St Jerome, the collapsing of time, as carried out on pilgrimage and in drama, art and literature (Holloway 2016, xvii-​xix). The encounter of the younger Margery with the older Julian, for instance as staging, in a parallel universe, Mary’s Visitation to Elizabeth. We know from The Book of Margery Kempe that Margery sought the advice of Julian concerning the “Discernment of Spirits” (as to whether her visions were from the Holy Spirit and thus conducive to charity to one’s fellow Christians, or whether they were from evil spirits and for self-​importance), in order to be reassured about her experiences because, as we know from the Book, Julian was considered an “expert”: in the Book’s words, “for the ankres was expert in swech thyngys and good cownsel cowd yevyn” (for the anchoress was expert in such things and could give good counsel). Being proved divinely inspired was never far from Margery’s mind: she paid a visit to Julian of Norwich, her contemporary, precisely to be reassured as to her contact with the divine. And canonization was perhaps not absent from the minds of the men who wrote down her book as she dictated it, shaping her reminiscences to fit into a long line of holy women (Dinshaw 2003). It seems that Julian’s contemplative authority parallels the clerical authority of the Vicar of St Stephen’s in Norwich Richard Caister and the White Friar William Southfield whom Margery visited at the same time. This is surely a motive explaining why Margery sought Julian’s counsel; there may be, however, additional reasons explaining why she sought the advice of Julian, and these may be related to her anxieties around trauma.

2. Women’s Soul Healing The Book of Margery Kempe attracted renewed scholarly interest at a conference held in Oxford in 2017. Particularly, Williams (2019: 14) discussed the conversation between Margery and Julian, their “Holy Dalyawnce” of several days, in her presentation citing recent dramas concerning Julian and Margery. Williams notes

The Soul a City: Margery meets Julian

35

that the Book positions the encounter of the two women most firmly within the genre of Margery’s encounters with religious authorities. According to Williams (2019: 12), Margery is interested in Julian not because of her femaleness but because of her visionary authority and she does not seek comfort and friendship but another seal of approval to add to her steadily-​growing list. However, we have also to take Margery’s anxieties and her need for validation into consideration: one that is specifically feminine, the other more linked to her general need of reassurance from authorities. We know from the Book that sex and childbearing was a constant preoccupation for Margery. After the trauma of the birth of her first child and her decision to change her lifestyle, she repeatedly tries to persuade her husband to live in chastity, though it is only after many years and children that her husband consents to the vow of chastity. In this difficult period she has foul visions of male organs and experiences even a temptation to have sex with an unknown man, who refuses her. We also know that she tended to consider childbearing a hindrance to her intention to dedicate her life entirely to Jesus. It may be that the counsel of a woman whose theology gave importance to Jesus not just as man (this an aspect very important for Margery, who, imitating Catherine of Siena and other saintly women, celebrates a mystical marriage with the divinity), but also as mother, as a feminine caring figure (Holloway 2019). A further source of trauma for Margery was the charge of heresy. Because of her peculiar way of life, she was on many occasions publicly scorned and even brought to trial before the Mayor of Leicester in 1417 under suspicion of being in connection with the Lollards (Book, ­chapter 48). The period between 1409, when archbishop Arundel promulgated his Constitutions, and 1415 was a phase of repression for Lollardy (Stavsky 2015). Literary texts provide consolation, soul healing, from the Book of Job, through Boethius’ Consolation of Philosophy, to Frankl’s Search for Meaning. as an alternative universe through which the reader can ‘sand box’ their own problems. Men, but especially women, use the figure of a woman as soul healer, Socrates’ Diotima, Boethius’ Philosophia, Dante’s Beatrice, Christine de Pizan’s Sibille. Such figures, in anthropology (Turner 1971) function as ‘adepts’. Julian, both in her Book of The Showing of Love, and in this recorded conversation in Margery’s Book, in Frankl’s and Turner’s terms, is such a Logotherapist and adept. This can be ‘Feminist Psychiatry’. Notwithstanding Margery’s many male supporters and counsellors, she also relies on a feminist spiritual network (Wilson 1997: 249–​273). Margery sought to copy Saint Birgitta of Sweden, like her a mother of many children, and like her a pilgrim to Compostela, Norway, Danzig, Aachen, Cologne, Rome, Bethlehem and Jerusalem. The writing of Birgitta’s Revelationes with the editing by Magister

36

Gabriella Del Lungo Camiciotti & Julia Bolton Holloway

Mathias and Bishop Hemming, by Bishop Hermit Alfonso of Jaen; then Catherine of Siena’s Dialogo, with William Flete, Raymond of Capua, Cristofano di Gano, and again, Alfonso of Jaen; then for Julian of Norwich’s Showings, with William Flete and Cardinal Adam Easton, is mirrored in the support given to Margery’s Book from multiple ecclesiasts and their writing it down, all four women profiting from the earlier model of Cardinal Jacques de Vitry’s scribal relationship with Marie d’Oignes (Holloway 2008: 217, etc; Allen –​Meech 1940 [1982: 153]). Examples from Birgitta’s own Revelationes and from later miracles reported in her Vita can demonstrate the effectiveness of her feminist credentials. In her Processus for sainthood it is told how at the birth of her eighth child, Cecilia, in Sweden, it was thought she would die but that Saint Mary, Jesus’ Mother, came and saved her life (Holloway et al. 1991). In Rome, when she was criticized for her poor Latin, she describes how she went to the Basilica of St Agnes, and prayed each day before the apse mosaic of the bejeweled martyred Roman girl, who, being a Latin native speaker, gave her excellent lessons, and also jewels from her crown for each insult she bore from men (Jørgensen 1954: 43–​44). Birgitta’s miracles often involved the healing of women, such as her maidservant, Katrina di Flandria, while she was alive, of the women Agnes de Comtessa, and the nun Francesca Sabelli while her body lay in state at Saint Lawrence in Panisperna, then later Elsebi Snara of Linköping, giving birth to a dead child, prays to St Birgitta and her babe becomes warm (Holloway 1991: 874–​875; 1997: 20). Margery was to meet a surviving maidservant of Birgitta’s in Rome, the English College (then the Hospital of St Thomas of Canterbury), where she stayed being adjacent to the Casa di Santa Brigida in Piazza Farnese (Allen –​Meech 1940 [1982: 89, 94–​95]; Holloway 1992: 209). While Julian’s Showing likewise copies Birgitta’s Revelationes and they narrate identically of women wanting to know the fate of loved ones and this not being revealed to them.

3. Scribalized Orality The oral reporting of mystical experiences on the part of pious lay women is a widespread practice of late medieval religiosity (Vauchez 1987); on a level, this can be interpreted as a form of psychotherapy, but primarily it reveals the search of ecclesiastical validation for their charisma, which could be given by the writing down of their memoirs, and was also a necessary step to initiate a process leading to the beatification of a saintly woman. The pervasiveness of orality in medieval culture is confirmed by the Book of Margery Kempe (Del Lungo 2000). Orality is its very texture, the scribes writing down her memoirs saying

The Soul a City: Margery meets Julian

37

that they are reporting only what Margery remembers and tells them verbally. So Margery authors (Del Lungo 2008) the content of her spiritual autobiography as the orally given word is still the privileged mode of credibility. In addition to that of the scribes, we hear many voices speaking in the Book as this is composed not only of narrative passages, but also, and predominantly, of reported speech, reported both directly and indirectly. We hear the voices of Christ and Margery conversing, even that of God and Saint Mary, or that of other saints, speaking to Margery. Indeed, a high proportion of the narration of Margery’s life is given in dialogic form (Del Lungo 2002, 2005), direct speech being mostly introduced by formulaic framing devices, which have a meta-​textual function by signalling to the reader/​listener that a certain speech event will follow; thus we can say that dialogue has a text organizing function, but it has also an expressive function. As stated by Tannen (1989), dialogue is an important source of emotions in orally reported speech as it creates interpersonal involvement, because meaning is dramatized, not stated by the narrator. It is also a frequent involvement device of late medieval religious culture which focused on stimuli for meditation to actively involve the reader/​listener (Love 2004). Sometimes, as in the case we present here, reported speech is given in a mixed mode, that is without explicitly marking the shift from narration to direct speech.

4. Paleography and Linguistics Editing literate Julian’s extant manuscripts, today in Paris’ Bibliothèque Nationale, in London’s British Library, and at Westminster, (Julian 2001), and in transcribing illiterate Margery’s reported dialogue with her from the Butler-​ Bowden manuscript, now British Library Additional 61823 manuscript, in “The Soul a City” on the Web, was to transcribe orally cadenced texts written with quills on parchment and on paper, with rubricated ampersands, initial letters and words in red for emphasis (here in bold), in their scribes’ differing dialects, from Norwich, “Dewch”, Lynn (the scribe Saltowes of Margery’s manuscript), Lincoln (the scribe of Julian’s Amherst manuscript), Midlands, some in the fifteenth century, others, later. These are texts influenced and “authorized” by male theologians giving spiritual direction to women, centred on the “Discernment of Spirits”, Hermit William Flete to Catherine of Siena, Magister Mathias and Bishop Hermit Alfonso of Jaen to Birgitta of Sweden, Cardinal Adam Easton to Julian of Norwich (Voaden 1999; Watt 1997b; Holloway 2008; Holloway –​Del Lungo 2017), all being distilled here by Margery as she reports Julian’s validating discourse. Our two women, and

38

Gabriella Del Lungo Camiciotti & Julia Bolton Holloway

with them, William Flete and Adam Easton, are all East Anglian and thus were speaking and writing in the dialects of Cambridge, Norwich and East Lynne, as identified in the LALME atlas (1986). They networked with individuals across Europe, through Latin, who are Swedish, Spanish and Italian, Margery’s daughter-​in-​law even being from the Hanseatic city of Danzig, today Gdansk. For Margery’s first scribe is noted as writing as if in a foreign tongue and with ill-​formed letters “Þe booke was so euel wretyn þat he cowd lytyl skyll þeron, for it was neiþyr good Englysch ne Dewch, ne þe lettyr was not schapyn ne formyd as oþer letters ben” (The book was so badly written that he could scarcely understand it, for it was neither in good English nor German, nor were the letters shaped and formed as other letters are). She could thus be the Danzig daughter-​ in-​law, coming from the Baltic city that revered Birgitta of Sweden, and who could have encouraged illiterate Margery, bereft of her son, to write a similar great book of Revelationes as had Birgitta as her Logotherapy and to make all the same pilgrimages Birgitta had made and described (Holloway 1992). Margery’s unique surviving manuscript in her Lynn Scandinavian-​influenced dialect comes to us from Yorkshire’s Carthusian Mount Grace Priory (Allen –​ Meech 1940 [1982: xi-​xlvi]), while Julian’s manuscripts are mostly copied out by Brigittine nuns of Syon Abbey in Richmond and in exile in the Middlesex dialect, though one late seventeenth manuscript written out by an exiled Benedictine nun carefully copies the dialect of the now lost autograph text Julian had written in her fourteenth-​century Norwich dialect (Greeson 2001: 627–​682). A further aid in placing their East Anglian vocabulary is a delightful Latin/​Middle English dictionary, the Promptorium Parvulourm, written by a Lynn anchorite, Galfridus Grammaticus (1908), who taught schoolboys.

5. Their Oral/​Scribal Conversation At the beginning Margery’s recalling of the event is reported as usual in the third person, but soon we shall hear the voice of Julian directly addressing Margery, sounds transcribed into letters, written out in secretary hand, some of which are rubricated for emphasis, here in bold, to be seen on folios and pages (Holloway 2016: 229–​245). GDL: Margery has her scribes tell us (fol. 21r) & þan ʃche was bodyn be owyr lord. for to gon to an ankres in þe ʃame Cyte which hyte Dame Jelyan. & ʃo ʃche dede & ʃchewyd hir þe grace þat god put in hir ʃowle of compunccyon contricyon ʃwetnesse & devocyon compaʃʃyon with holy meditacyon & hy contemplacyon. & ful many

The Soul a City: Margery meets Julian

39

holy ʃpechys & dalyawns. þat owyr Lord ʃpak to hir ʃowle. and many wondirful reuelacyons whech ʃche ʃchewyd to þe ankres to wetyn yf þer were any deceyte in hem, for þe ankres was expert in ʃwech thynges & good cownʃel cowd ʒeuyn. ‘And then she was told by our Lord, to go to an anchoress in the same city called Dame Julian. And so she did and showed her the grace that God put in her soul of compunction, contrition, sweetness and devotion, compassion with holy meditation and high contemplation. And many holy speeches and dalliance that our Lord spoke to her soul, and many wonderful revelations which she showed to the anchoress to know if there were any deceit in them, for the anchoress was expert in such things and could give good counsel.’

Margery and Julian’s conversation continues: Þe ankres, heryng þe meruelyows goodnes of owyr lord, hyly thankyd god. with al hir hert. for þys viʃytacyon cown∫elyng þis creature to be obedyent. to þe wyl of owyr lord god & fulfyllyn with al hir myghtys. whateuer he put in hir ʃowle yf it wer not ageyn þe wor shep of god & profyte of hir euyne criʃten, ‘The anchoress, hearing of the marvelous goodness of our Lord, highly thanked God, with all her heart for this visit, counselling this creature to be obedient to the will of our Lord God and to fulfill with all her might whatever he put in her soul, if it were not against the worship of God and profit of her even Christian.’

JBH: Julian continues in her conversation with Margery, which is now reported in direct speech:     for yf it were þan it were nowt þe mevyng of a good ʃpyryte but raþer of an euyl ʃpyrit. Þe holy goʃt meuyth neuyer a þing a-​geyn charite &, yf he dede he were contraryows to hys owyn ʃelf for he is al charite. Al∫o he meuyth a ʃowle to al chaʃtenesse. for chaʃt leuars be clepyd þe temple of þe holy goʃt . & þe holy goʃt makyth a ʃowle ʃtabyl & ʃtedfast in þe rygth feyth & þe rygth beleue. And a dubbyl man in ʃowle is euer vnʃtabyl. & vnʃtedfaʃt in al hys weys. He þat is euermor dowtyng. is lyke to þe flood of þe ʃee. þe wheche is mevyd & born a-​bowte with þe wynd, & þat man is not lyche to receyuen þe ʒyftys of god. What creature þat hath þes tokenys he mu∫te ʃtedfaʃtlych belevyn þat þe holy goʃt dwellyth in hys ʃowle.

40

Gabriella Del Lungo Camiciotti & Julia Bolton Holloway And mech more whan God viʃyteth a creature wyth terys of contriʃyon deuoʃyon er compaʃʃyon. he may & owyth to leuyn þat þe holy goʃt is in hys ʃowle. ‘For if it were then it were not the moving of a good spirit but rather of an evil spirit. The Holy Ghost never impels a thing against charity. And if he did he were against himself, for he is all charity. Also he moves a soul to all chastity, for chaste lovers are called the temple of the Holy Ghost [1 Cor. 6.19]. And the Holy Ghost makes a soul stable and steadfast in the true faith and right belief. And a man who is double in soul is always unstable and unsteadfast in his ways. He who is always doubting is like the flood of the sea, which is moved and born about by the wind, and that man is not likely to receive the gifts of God. Who has these tokens must [fol.21v] steadfastly believe that the Holy Ghost dwells in his soul. And much more when God visits a creature with tears of contrition, devotion or compassion, he may and ought to believe that the Holy Ghost is in his soul. And much more when God visits a creature with tears of contrition, devotion or compassion, he may and ought to believe that the Holy Ghost is in his soul.’

Julian next is reported as citing her authorities, Paul and Jerome, to Margery, who perhaps misremembers one of them: Seynt Powyl ʃeyth þat þe Holy Goʃst Aʃkyth for vs with morningges & wepynges vnʃpekable. þat is to ʃeyn he makyth vs to aʃkyn & preyn with mornyngges & wepynges ʃo plentyvowʃly. þat þe terys may not be nowmeryd. Ther may non euyl ʃpyrit ʒeuyn þes tokenys, for Sanctum Jerom ʃeyth þat terys turmentyn more the Debylle þan don the peynes of Helle. ‘St Paul says that the Holy Ghost implores us with unspeakable mourning and weeping. That is to say he makes us to ask and pray with such plenteous mourning and weeping that the tears may not be counted [Romans 8.26]. There may be no evil spirit given these tokens, for St Jerome says that tears torment the devil more than do the pains of hell.’

Julian next discusses evil: god & þe deuyl ben euermor contraryows & thei xal neuer dwellyn togedyr in on place. & þe devyl hath no powyr in a mannys ʃowle. Holy wryt ʃeyth þat þe ʃowle of a rytful man is the ʃete/​seet of God. & ʃo I truʃt, ʃyʃter, þat ʒe ben. ‘God and the devil are always contrary and they shall never dwell together in one place. And the devil has no power in a man’s soul. Holy Scriptures say that the soul of a rightful man is the city/​seat of God. And so I trust, sister, that you are.’

With that last comment, “& ʃo I truʃt, ʃyʃter, þat ʒe ben”, we realise that we certainly are listening to reported speech and that Dame Julian addressed Dame Margery, her “evyn christen” ‘even-​Christian’, even as her “Sister”.

The Soul a City: Margery meets Julian

41

6. Conclusion To study the two women we thus need theology, psychiatry, anthropology, philology, paleography, codicology, linguistics –​and also women studies for their soul-​healing through books. In this way we can restage their encounter and in their voices hear again those of many others, of both genders, bass and treble, dialoguing across all of Europe, from both before and after that Visitation to the aged staid Julian in her Norwich graveyard by the younger ecstatic Margery clad in white traversing the then known world on pilgrimage.

References Allen, Hope Emily –​Sanford Brown Meech (eds.). 1940. Margery Kempe. The Book of Margery Kempe. EETS 212. London: Oxford University Press. Reprinted 1982. Beale, Jane (ed.). 2019. Illuminating Jesus in the Middle Ages. Leiden, Boston: Brill. Betten, Anne –​Dannerer, Monika (eds.). 2005. Dialogue Analyse IX /​Dialogue Analysis –​Dialogue in Literature and the Media. Tübingen: Max Niemeyer. Birgitta of Sweden. 1997-​. Revelationes. http://​www.umilta.net/​Birgitta.html [accessed 2 July 2020]. Blunt, John H. (ed.). 1873. The Myroure of Oure Ladye. (EETS e.s. 19). London: Oxford University Press. Del Lungo Camiciotti, Gabriella. 2000. “Orality and written texts: the representation of discourse in the Book of Margery Kempe”. In: Di Martino, Gabriella –​ Maria Lima (eds.): 143–​157. Del Lungo Camiciotti, Gabriella. 2005. “Metanarrative frame and evaluation in late medieval saints’ lives”. In: Betten, Anne –​Dannerer, Monika (eds.): 269–​278. Del Lungo Camiciotti, Gabriella. 2008. “Self-​disclosure as co-​constructed discursive activity in the Book of Margery Kempe”. In: Douthwaite, John –​ Domenico Pezzini (eds.): 18–​29. Di Martino, Gabriella –​Maria Lima (eds.). 2000. English Diachronic Pragmatics. Napoli: CUEN. Dinshaw, Carolyn. 2003. “Margery Kempe”. In: Dinshaw, Carolyn –​David Wallace (eds.): 222–​239. Dinshaw, Carolyn –​David Wallace (eds.). 2003. The Cambridge Companion to Medieval Women’s Writing. Cambridge: CUP.

42

Gabriella Del Lungo Camiciotti & Julia Bolton Holloway

Douthwaite, John –​Domenico Pezzini (eds.). 2008. Words in action: diachronic and synchronic approaches to English discourse. Studies in Honour of Ermanno Barisone. Genova: Ecig. Frankl, Viktor. 2006. Man’s Search for Meaning: An Introduction to Logotherapy. Boston, MA: Beacon Press. Greeson, Hoyt S. 2001. “Glossary to the British Library Sloane 2499 Manuscript”. In: Reynolds, A.M. –​J. Bolton Holloway (eds.): 627–​682. Holloway, Julia Bolton. 1992. “Bride, Margery, Julian and Alice: Bridget of Sweden’s Textual Community in Medieval England”. In: McEntire, Sandra (ed.): 203–​221. Holloway, Julia Bolton. 2000. Saint Bride and her Book: Birgitta of Sweden’s Revelations. Woodbridge, Suffolk: Brewer. Holloway, Julia Bolton. 2008. Anchoress and Cardinal: Julian of Norwich and Cardinal Adam Easton, O.S.B. (Analecta Cartusiana 35–​20). Salzburg: Institut für Anglistik und Amerikanistik Universität. Holloway, Julia Bolton. 2016. Julian Among the Books: Julian of Norwich’s Theological Library. Newcastle-​upon-​Tyne: Cambridge Scholars Publishing. Holloway, Julia Bolton. 2017. Christine de Pizan/​Cristina da Pizzano. Le Chemin de Longs Etudes/​Il Cammin di Lungo Studio. Trans. Ester Zago. Parallel text, French/​Italian. De strata francigena. Florence: Centro Studi Romei. Holloway, Julia Bolton. 2019. “Jesus as ‘Mother’ ”. In: Beale, Jane (ed.): 291–​309. Holloway, Julia Bolton et al. 1993. “A Bridgettine Document from the Florentine Paradiso written at Vadstena, 1397, and its Context”. In: Santa Brigida profeta dei tempi nuovi: Saint Birgitta, Prophetess of New Ages: Proceedings of the International Study Meeting. Rome: Casa Generalizia Suore Santa Brigida: 860–​900. Holloway, Julia Bolton –​Gabriella Del Lungo Camiciotti (eds.). 2017. Mary’s Dowry; An Anthology of Pilgrim and Contemplative Writings: La Dote di Maria: Antologie di Testi di Pellegrine e Contemplativi. Analecta Cartusiana 35. Salzburg: Institut für Anglistik und Amerikanistik Universität Salzburg. Jørgensen, Johannes.1954. Saint Bridget of Sweden. Trans. Ingeborg Lund. London: Longmans Green. LALME = McIntosh, Angus –​Michael L. Samuels –​Michael Benskin (eds.) (with the assistance of Margaret Laing and Keith Williamson). 1986. A linguistic atlas of Late Mediaeval English. 4 vols. Aberdeen: Aberdeen University Press. http ://​www.lel.ed.ac.uk/​ihd/​elalme/​elalme.htm [accessed 2 July 2020]. Lawes, Richard. 2000. “Psychological Disorder and the Autobiographical Impulse in Julian of Norwich, Margery Kempe and Thomas Hoccleve”. In: Reveney, Denis –​Christiana Whitehead (eds): 217–​239.

The Soul a City: Margery meets Julian

43

Mayhew, Anthony L. (ed.). 1908. Galfridus Grammaticus. The Promptorium Parvulorum sive Clericorum: The First English-​Latin Dictonary. (EETS 102). London: Kegan Paul, Trench, Trübner. McEntire, Sandra (ed.). 1992. Margery Kempe: A Book of Essays. New York: Garland. Papebroeck, Daniel (ed.). 1867. “Vita Maria Oigniacensi in Naurcensis Belgii diocesi per Jacobum de Vitriaco”. Acta sanctorum. Ed. novissima 5: 542–​572. Pezzini, Domenico. 2015. Una rivelazione dell’Amore . Milano: Ancora. Reveney, Denis –​Christiana Whitehead (eds). 2000. Writing Religious Women: Female Textual Practices in Late Medieval England. Cardiff: University of Wales Press. Reynolds, Anna Maria –​Julia Bolton Holloway (eds). 2001. Julian of Norwich. Showing of Love. Florence: SISMEL Edizioni del Galluzzo. Stavsky, Jonathan. 2015. “’As the Lily among the thorns’: Daniel 13 in the Writings of John Wyclif and his Followers”. Viator 46: 249–​274. Sargent, Michael G. (ed.). 2004. Nicholas Love. The Mirror of the Blessed Life of Jesus Christ. Exeter: University of Exeter Press. Tannen, Deborah. 1989. Talking Voices, Repetition, Dialogue, and Imagery in Conversational Discourse. Cambridge: Cambridge University Press. Turner, Victor. 1971. The Ritual Process: Structure and Anti-​ Structure. Ithaca: Cornell University Press. Vauchez, André. 1987. Les laïcs au Moyen Ages. Pratiques et Expériences Religieuses. Paris: Les éditions du Cerf. Voaden, Rosalynd. 1999. God’s Words, Women’s Voices. The Discernment of Spirits in the Writing of Late Medieval Women Visionaries. York: York Medieval Press. Watt, Diane (ed.). 1997a. Medieval Women in their Communities. Cardiff: University of Wales Press. Watt, Diane. 1997b. Secretaries of God: Women Prophets in Late Medieval and Early Modern England. Cambridge: Brewer. Williams, Tara. 2019. “Revisiting Margery and Julian’s Holy Dalyawns”. Presented at Margery Kempe Studies in the 21st Century. University College, Oxford, 5th -​ 7th April 2018. Wilson, Janet. 1997. “Communities of Dissent: The Secular and Ecclesiastical Communities of Margery Kempe’s Book”. In Watt, Diane (ed.): 249–​273. Windeatt, Barry (ed.). 2000. Margery Kempe. The Book of Margery Kempe. Harlow, Essex: Longman. Wålin, Lars –​Margarete Andersson-​Schmitt (eds.). 1990. Magister Mathias Copia exemplorum. (Studia Latini Upsaliensis 2). Uppsala: Universität Uppsala.

Omar Khalaf

Patronage, Print and the Education of the Gentry in Late Medieval England: The Case of Earl Rivers’s Dicts and Sayings of the Philosophers Abstract: This essay focuses on the production and circulation of Earl River’s Dicts and Sayings of the Philosophers. I argue that Rivers’s patronage of William Caxton during the printer’s early years at Westminster was part of a precise cultural plan, which entailed the dissemination of moral and educational literature outside the court. As revealed by a preliminary investigation of the ownership of incunables and manuscript witnesses, the readership of Rivers’s Dicts is to be found in the gentry –​a social class interested in books as both receptacles of typically courtly literary models and material symbols of a newly-​ acquired power. Keywords: Earl Rivers, Caxton, Dicts and Sayings of the Philosophers, education, gentry

1. Earl Rivers and the Dicts and Sayings of the Philosophers This is what King Edward IV disposed for the education of his heir Edward in 1473: [We wille that] noe man sytt at his boarde but such as shal be thought by the discretyon of the [...] Erle Ryvers, and that then be reade before him such noble storyes as behoveth to a Prynce to understande; and knowe that the comunicatyon at all tymes in his presence be of vertu, honor, cunynge, wisdom, and deedes of worshippe, and of nothing that should move or styrre him to vyces. (Nichols 1790: 28)

This ordinance was delivered to the Council of the Prince of Wales when Edward, at the age of three, was sent to Ludlow and put in charge of the western part of the kingdom with the support of John Alcock, bishop of Rochester, and the “ryght trusty and wel beloved” –​as defined in the document –​Anthony Woodville, Earl Rivers. Brother of Queen Elizabeth Woodville, Rivers is renowned among historians for his involvement in the tragic incidents that followed Edward IV’s death and brought his brother Richard III to the throne in 1483. Nevertheless, one of Rivers’s greatest accomplishments is his patronage of William Caxton during the printer’s early years in Westminster. Rivers’s most important work, the Dicts and Sayings of the Philosophers (henceforth Dicts) was the first book

46

Omar Khalaf

ever printed in England to bear its date of publication –​18 November 1477. The print of this text marked the beginning of a fruitful collaboration between Rivers and Caxton, which favoured the issue of two other writings by the earl: the Moral Proverbs, a translation of Christine de Pizan’s Proverbes Moraulx in 1478 and the Cordyal, an English rendering of Jean Miélot’s Les quatre choses derrenieres in 1479. The source of the Dicts is a substantial compilation of proverbs and maxims known as Kitab Mokhtâr el-​Hikam Wa-​Mahasin Al-​Kalim (The book of the choicest maxims and best proverbs) written in the eleventh century by the Arab Mubachschir ben Fatik, which widely circulated in medieval Europe by means of numerous translations and re-​elaborations. The thirteenth-​century Castilian translation of the Kitab (the Bocados de oro) served as the source of the Latin version, produced by Giovanni da Procida for Emperor Frederick II in the second half of the same century and titled Liber Philosophorum Moralium Antiquorum. Between the end of the fourteenth century and the beginning of the fifteenth, the Liber was turned into French by Guillaume de Tignonville, provost of Paris. This version, known as Dits moraulx des philosophes, caused three different English translations. The first was authored by Stephen Scrope around 1450 and revised by William Worcester in 1472. The second, anonymous, was produced presumably shortly after Scrope’s. Rivers’s translation is the latest. In the Prologue Rivers tells that he acquired the French text in July 1473 during his pilgrimage to Santiago de Compostela. Appreciating the edifying contents of the book and (possibly) unaware of the previous English versions (see Khalaf 2017), he resolved to translate it for the benefit of his nephew Edward, the Prince of Wales, in 1473. Consequently, Rivers must have completed his translation between 1473 and 1477. Rivers’s text, which follows its French source rather closely,1 contains the precepts of diverse ­figures –​mostly Greek philosophers, along with other historical or fictitious characters: Sedechias, Hermes, Gac, Zalquinus, Homer, Salon, Sabyon, Ippocras, Pythagoras, Diogenes, Socrates, Plato, Aristotle, Alexander the Great, Ptolemy, Assaron, Legmon, Anese, Sacdarge, Thesile, Gregory the Great, Galen. A final section collects a miscellany of sayings attributed to Anaxagoras, Aristophanes, Philip of Macedon, and some other anonymous wisemen. Each chapter is introduced by a short biography of the character followed by a collection of proverbs, sometimes inserted in an anecdotical context.

1 The only critical edition of the Ditz moraulx available is found in Eder (1915).

The Case of Earl Rivers’s Dicts and Sayings of the Philosophers

47

Although the long series of precepts contained in this text makes it quite uninteresting to a modern readership,2 the widespread success of gnomic and sentential literature not only in England, but in all Europe, is the key to understand the fortune of works like the Dicts. This genre, which flourished in England since the twelfth century, can be identified in two categories: the collections of parent-​to-​ child teachings, which regard the intimate sphere of family life and are meant for the salvation of the soul –​Myne Awen Dere Sone and How þe Gode Wyfe Tauht hyr Douȝter are two remarkable ­examples –​and the florilegia of sayings attributed to authorities of the past. In England, the most successful of these is the alliterative poem titled Proverbs of Alfred, written in the thirteenth century. Besides individual and filial education, moral literature was widely employed for the education of future sovereigns along with chivalric romances and specula principum (see Orme 1984; 2005). The most renowned were John of Salisbury’s Policraticus (1159), Gerald of Wales’s Liber de principis instructione (1193) and De instructione principis (date uncertain), and Thomas Hoccleve’s De regimine principum (beginning of the fifteenth century). It was Rivers who included the Dicts in this genre, as he deemed the contents of the book suitable for the education of the prince, in accordance with what had been disposed by his kingly brother-​in-​law. In Rivers’s Prologue we read: After suche season as it lysted the kynges grace comaunde me to gyue myn attendance vppon my lord the Prince and that I was in hys seruyse, whan I had leyser I loked vpon the sayd booke and at the last concluded in my self to translate it in thenglyssh tonge [...] thynkyng also ful necessary to my said lord the undestandyng therof ([a]‌3r) (Blades 1877: 3).

When Rivers decided to undertake his pilgrimage in Spain, Prince Edward was three years of age. In the fall of that same year the King issued the ordinances, which were conceived as a guide to his son’s everyday life. All the planned activities had to follow a precise schedule: worship, breakfast, learning, lunch, learning again, sports, evening worship, leisure. Bedtime at eight in the evening (see Nichols 1790: 28). Although education occupied a significant part of Edward’s day, scholars are quite dubious on the actual level of instruction the prince reached, especially if compared with his Tudor successors. For instance, he was given a teacher in grammar (i.e., Latin) since the age of four, but it is impossible to determine whether the prince was actually proficient in this language if in the 2 “Dicts and Sayings is by no means a work of high art –​though, to be fair, it was never intended as such. Instead it is a disorderly, repetitive, almost overwhelming heap of recorded wisdom” (Sutton 2006: 10).

48

Omar Khalaf

updated version of the ordinances issued just before his death in 1483 the king decided to remove it from the activities planned for his son (Orme 1984: 77). Aysha Pollnitz questions the records of the chroniclers of that time –​depicted as “extremely unctuous” by the scholar herself (2015: 26) –​among whom the Italian Domenico Mancini. In his De occupatione regni Anglie per Riccardum Tercium, Mancini describes the prince as well educated in literary subjects. He could read and understand works in prose and verse (1969, 92–​3), but no mention is found to the languages he mastered. Rivers’s educational plan, on the other hand, seems to be focused on literary matters and the study of works in English. Also the other translations by Rivers –​the Moral Proverbs and the Cordyal –​ were most likely addressed to Edward, who was perhaps too young at the time to understand texts in French. In 1477, the same year of the publication of the Dicts, Caxton dedicated his History of Jason (a translation of Raoul Lefevere’s Histoire de Jason), hoping that it might help Edward to “begynne to lerne to rede Englissh”.3 Consequently, it is possible that Rivers’s programme was based on the reading, and at least in the early years of Edward’s life, on the listening of works written in his mother tongue. Only at a certain age could the prince read and understand works in French. According to Pollnitz, Edward’s literary tastes can be considered rather conventional: “He owned epic tales in French, Burgundian stories of chivalry and anti-​French literature” (2015: 27). This situation of multilingualism in Edward’s reading practice is well exemplified by Oxford, Bodleian Library, MS Bodley 264, a compilation of French texts on the life and deeds of Alexander the Great (including the Roman d’Alexandre), in which an anonymous compiler interpolated the only copy of the English alliterative poem Alexander and Dindimus. Generally considered one of the most lavish witnesses of Flemish book decoration, this book was part of the Woodvilles’ library since 1466. As it combines chivalric entertainment with the moral teachings contained in the English interpolation, this codex was possibly one of the tools employed by Rivers for the education of his nephew (Khalaf 2011). Apart from being the first dated book published by Caxton in England, the Dicts also holds the primacy for the number of editions issued during the printer’s early years at Westminster. The first, which was followed by another in ca. 1480

3 See Crotch (1928: 34). Weinberg (2005: 53), argues: “It seems highly likely that this text [the History of Jason] was part of the educational program set out for the prince, and Woodville could have been the intermediary in prompting Caxton’s translation”.

The Case of Earl Rivers’s Dicts and Sayings of the Philosophers

49

and yet another in 1489,4 raises some important questions over the circulation of the text through this new technology and the readership it was meant to address. According to the Prologue, Rivers’s translation had been conceived to serve a private purpose, related to the education of the prince. Why did then Rivers decide to favour a large-​scale print production of this text? Although largely neglected by scholars, this matter paves the way for interesting reflections over the cultural policy implemented by Rivers and, not secondarily, the impact of the printing press in the dissemination of literacy outside the sphere of the court.

2. The Dicts, Rivers’s Patronage, and the Education of the Gentry Recently, the issue related to the education of the gentry and the new role it was called to play in the management of the kingdom since the beginning of the fifteenth century has become the object of considerable scholarly attention. Particular stress has been put on the increasing importance this social class was acquiring through the appointment of many of its members to positions that had been previously reserved to the highest ranks of nobility. Edward IV himself promoted this change, as he was seeking larger support against the nobles who had opposed him siding with Henry VI during the War of the Roses.5 Moreover, the progressive abandonment of the typically feudal economic model which was fully realised in the mid-​fifteenth century pushed numerous lesser nobles to undertake urban trade activities (Truelove 2005: 85–​86 and Orme 2005: 63). The access to local and central administration and the management of economic investments were the major causes for the gentry’s growing need of literacy.6 This

4 One more edition was published by Wynkyn De Worde in 1528. While Rivers’s prologue and Caxton’s epilogue were retained, De Worde added woodcuts at the beginning of each chapter. 5 The Woodvilles benefited greatly from this situation. After Elizabeth’s marriage with Edward IV in 1464, her father Richard was created Earl Rivers and appointed Constable of England. Anthony, her elder brother, succeeded as earl and became Governor to the Prince of Wales. This charge had also political relevance, as the family acquired control of the heir and could have influence all his future choices as king. At Edward IV’s death his brother Richard Duke of Gloucester seized the power taking custody of the prince and had Rivers executed. The same year he rose to the throne as Richard III. His possible involvement in the death of Prince Edward and his brother –​the famous “Princes in the Tower” –​is still matter of debate. 6 See Clanchy (1993: 13). The question regarding the level of literacy in England between the fifteenth and the sixteenth century is still matter of debate. Cressy (1980: 176–​177) assumes that about 10 % of men and 1 % of women in England were at least able to

50

Omar Khalaf

situation determined an increase in the demand –​and consequently in the production –​of books, which, apart from providing material supports for reading and instruction, started being conceived as status symbols, tangible objects of the owner’s wealth and social standing.7 I do not share Alison Truelove’s pessimistic view when she states: although we have some idea of the kind of education some members of the gentry may have received, and evidence of their ownership and use of literary manuscripts and books, this is unreliable evidence of their particular literary skills. Books, for example, were regarded as valuable prestige items, the ownership of such not necessarily denoting an ability to read them (Truelove 2005: 88).

Evidence against this position is provided, for instance, by the rich epistolary exchanges occurred between various members of the Paston, Stonor and Haute families, which, apart from providing precious information to historians, also witness the level of literacy the gentry –​or at least part of it –​had reached. This “group of consumers and shapers of literary culture” (Johnston 2014: 27) started acquiring even greater relevance than the higher ranks of nobility in the material production of literature. The gentry’s new social and cultural achievement justifies their need to aim at a level of education that had been previously reserved to the court circle and explains the wide circulation of romances and edifying works8 like the above-​mentioned proverbial literature and of several specula principum. In this regard,9 Orme argues: The medieval gentry [...] read about education in didactic literature, whose purpose was to instruct. A large category of writings, nowadays known as ‘mirrors for princes’ described how kings and nobles should be educated and what they should learn about. One of the most popular was the Secretum Secretorum [...]. Not only royalty read it; its many versions and manuscripts show that it was popular among the nobility and gentry (Orme 2005: 66).

sign. More optimistically, Lester maintains that since the beginning of the fifteenth century people who could read, write, and profit from literary works could be found at all social levels. Cf. Bennett (1970), Parkes (1991), and Trapp (1999). 7 See Youngs (2005: 119): “The book was an important accessory to the gentry lifestyle. Despite the commercialisation of the manuscript/​book trade in the fifteenth century, the book remained a luxury good and a symbol of affluence”. 8 The circulation of romances within this social class has been thoroughly investigated. See inter alios Riddy (1987), Hudson (1984), Putter (2000), and Johnston (2014). On the other hand, the (re)production and use of didactic works still requires further consideration. For general remarks see Keiser (1999) and Salter (2012: 137–​176). 9 See also Radulescu (2005: 101 and 109).

The Case of Earl Rivers’s Dicts and Sayings of the Philosophers

51

Undoubtedly, the proliferation of the Dicts in printed form played a primary role in this context. Twenty incunables of the first two editions are extant –​the more relevant here, as they were issued during Rivers’s lifetime –​which is a remarkable number if compared with other publications by Caxton in his first years of activity. The Dicts, in fact, are second only to the first edition of the Canterbury Tales, published in 1476–​77, which counts a total of seventy-​seven extant copies; yet, while this work is generally considered Caxton’s first great investment in book production, the publication of the Dicts was the outcome of the active patronage of Rivers, who personally commissioned its revision and print. Caxton made it clear in the Epilogue, in which his patron is urged to take the labour of thenpryntyng in gre & thanke, whiche gladly haue don my dyligence in thaccompysshyng of his desire and commandement, in whiche I am bounden so to do for the good reward that I haue resseyuyd of his sayd lordshyp ([k]‌6v) (Crotch 1928: 30).

Rivers’s activity as translator and his sponsorship of Caxton fulfilled the demand of literacy manifested by the gentry. Providing access to a book expressly produced for the education of the Prince of Wales, Rivers and Caxton gave a crucial contribution in providing the future managing class of the kingdom with useful teachings and moral guidance. Moreover, the exchange of books determined the creation of cultural and political connections between families and individuals. Scholars have focused on the circles that gravitated around John Fastolf, the Pastons and the Hautes, and on the role they played in the creation of connections among the gentry.10 The development of such networks favoured the birth and consolidation of spheres of influence which involved the authors, the producers, and the readers in a dynamic relationship. Rivers played a paramount role in this context, as he exploited the potentialities offered by Caxton’s printing press and marketing skills to impose his literary authority favouring –​and probably managing himself –​the production and dissemination of his works.

3. The Circulation of the Dicts: Some Examples Although a full investigation of the Dicts has not yet been attempted, it is possible to speculate on the circulation of its witnesses starting from the studies on book production and ownership in the fifteenth century. In her outline of private owners of printed books in medieval Britain, Margaret Lane Ford focuses on the incunable as a medium of social and political connections between the 10 See Youngs (2005). For Fastolf, the Pastons, and the Hautes see, respectively, Hughes (1992), Richmond (2000), and Fleming (1987).

52

Omar Khalaf

gentry and the nobility. According to her, these relationships become particularly visible considering the ownership of the first incunables printed by Caxton, as many of these owners “were related or acquainted” (1999: 213). Probably, Rivers’s social standing and his role of Governor of the Prince of Wales contributed to the success of the Dicts and prompted its circulation, not only because of the edifying contents of the book, but –​and perhaps mostly –​for the prestige that the ownership of one its copies might entail. Prestigious patronage certainly influenced the ownership of the early English books, both directly through distribution of sponsored books, and indirectly through prestige placed on such ownership. [...] Books owned by the gentry-​merchant classes were not directly necessary to conducting their affairs, but, as devotional aids and works for edification as well as entertainment, Caxton’s books were necessary to their life-​styles (Ford 1999: 218).

A preliminary investigation of some copies of the Dicts confirms the social value of book ownership in the establishment or consolidation of personal or family connections. Ford herself reports an example related to a copy of the second edition: London, British Library, C. 10b. 2 belonged to the St. John family, whose member was Eleanor, the wife of Thomas Grey, one of the children that Queen Elizabeth Woodville had had with her first husband (1999: 314). A copy of the Dicts also appears in the library of James Morice, clerk to Lady Margaret Beaufort, who mentions it in a list he wrote himself in his copy of the 1481 edition of William Worcester’s translation of De senectute, now held at Cambridge University Library.11 This small inventory, which includes a Life of Christ, the Stans puer ad mensam (probably in translation), Lydgate’s Temple of Glass and the Canterbury Tales, is one of the best examples of what a gentleman of the rank of Morice should read. Margaret Beaufort was one of the protagonists of the events that followed Rivers’s death and the rise to power of Richard III. Mother of Henry Tudor, she was one of the most zealous supporters of her son in his claim for the English throne. A connection between her and the Woodvilles can be seen in her negotiation with Elizabeth Woodville for the marriage of Henry to Elizabeth of York, the daughter of Edward IV. Also the extant manuscripts provide valuable clues on the circulation of the Dicts in late fifteenth century; moreover, they account for the capillarity of the dissemination of Rivers’s work not only thanks to Caxton’s prints, but also through manuscript copies of his editions.12 The reproduction of the text in handwritten 1 1 Cambridge University Library, Inc 3.J.i.i, [a]‌1v. see Oates (1960) and Jayne (1956: 93). 12 London, Lambeth Palace Library MS 265, Chicago, Vault folio Case MS 36 (olim Newberry Library MS f.36 Ry 20) fols. 208r-​241r and New York, Columbia University

The Case of Earl Rivers’s Dicts and Sayings of the Philosophers

53

copies suggests that the fortune of the work exceeded Rivers’s and Caxton’s most optimistic expectations. Moreover, the codicological context in which each witness is found helps to understand the function attributed to the text as well as its owner(s). In fact, the manuscripts can be distinguished in two types: those containing full-​text copies of the Dicts and those presenting epitomised versions or fragments. Most full-​text copies are full-​text copies de facto, as they reproduce faithfully all the elements of their source, including the references to its print and publication found in Caxton’s Epilogue. The most telling example is London, Lambeth Palace Library, MS 265. Produced in 1477 as a Christmas gift from Rivers to the royal family, it contains an illustration representing the earl kneeling before the royal family and handing the book to the king. As this copy was produced only few weeks after the issue of Caxton’s first edition, we can assume that Rivers himself still valued a manuscript copy of his text a more convenient object to be bestowed at court than a printed book. The copy held in London, British Library MS Additional 22718 is equally meaningful. As I wrote elsewhere (Khalaf forthcoming), the manuscript contains some notes written by its former owner Lord Walter C. Trevelyan, who bequeathed it to the British Library in 1859. The flyleaves record numerous birth dates of members of the Hill (or Hyll) family of Spaxton, Somerset. The most interesting entry is Maud Hill, born in 1505. Her mother Alice was a member of the Stourtons, another Somerset gentry family that was related to the court thanks to the close relation between Alice’s uncle William and John de la Pole, Duke of Suffolk and husband of Edward IV’s sister Elizabeth, Elizabeth Woodville’s sister-​in-​law. The abridgements and the fragments of the Dicts demonstrate that the work was also valued for its contents. New York, University of Columbia Library MS Plimpton 259 contains a version of the Dicts interpolated in a typically secular context of use. This manuscript, which is a classical example of commonplace book, has not been the object of detailed studies except for the mathematical treatise it contains (Acker 1993). Besides them and various sapiential and gnomic texts (an unidentified compilation of proverbs in Latin titled Dictum philosophorum, maxims from various authors, the poem The Words of a Good Library, MS Plimpton 259 are derived from Caxton’s first edition; Dublin, Trinity College MS 213 fols. 70v-​72r, and New York, Pierpont Morgan Library MS B.11 are considered copies of the second edition. See Blake (1991: 295–​296). I have recently demonstrated that the witnesses contained in London, British Library MS Add. 22718 and in Add. 60577, fols. 38r-​44v do not derive directly from one of Caxton’s prints, but from a now lost text containing most of the innovations found in the second edition. See Khalaf (2018).

54

Omar Khalaf

Horse to his Master, adages in Latin and English and Saint Bernard’s De contemptu mundi),13 on fols. 54r–​61v14 is found a very condensed epitome of the Dicts which includes a choice of some of the philosophers found in Rivers’s text (we find, in the spelling of the copyist, Hermes, Sabyon, Ypocras, Pytagoras, Dyogenes, Socrates, Platon, Alyzaundyr, Ptholome, Legmon, Assaron, Anes, Pthesylim, Galeon, and Protege). The first owner of the book was a certain Robert Gottes (or Gottys) of Little Ryburgh, Norfolk, who left some records on fols. 11v–​15v and 44r. Some more details found by Acker (1993: 78–​79 n.7) help identify Gottes as a contemporary of Rivers, active in the acquisition of lands in the Norfolk area in 1475–​1477. Other fragments show that the text also circulated in religious environments. One of them is London, British Library, MS Add. 60577, which contains the section of the Dicts concerning Hermes15 along with some other moralizing works copied for the monks of St. Swithun’s Abbey. Closely connected with the priory of Winchester, St. Swithun was under the direct supervision of the bishop and celebrated humanist William Waynflete. A point of connection between Rivers and Waynflete can be established in Eton College. Rivers was one of the patrons of the institution, as he secured Eton a grant of a tenement in London around 1475; for this reason, a mass was sung everyday for him and his family, which had to be preceded by sixty strokes (Lyte 1911: 70). Waynflete had been provost of the college between 1442 and 1447 and he remained involved in its development until his death in 1486 (Lyte 1911). Following his educational policy, Rivers may have sent a printed copy of his Dicts to the bishop, who made it available to the monks for study and reproduction. The codicological context in which this fragment is found is worth of notice:16 extant on fols. 38r–​44v, it was placed by the copyist next to Lydgate’s translation of the Secretum Secretorum. Possibly conceived as an integration to Lydgate’s work by the compiler of the codex, here the Dicts acquire the same dignity as the most renowned mirror for princes that circulated in the Middle Ages.

1 3 14 15 16

See Ives (1942: 45) for a revision of De Ricci’s census (1935–​1940). The actual foliation does not reconcile with that provided in Bühler (1948). For a study of the witness see Khalaf (2018). The manuscript has been thoroughly studied by Wilson and Fenlon (1981) in the introduction to their facsimile edition.

The Case of Earl Rivers’s Dicts and Sayings of the Philosophers

55

4. Conclusions A full account of the circulation of the Dicts in late fifteenth-​and early sixteenth-​ century England requires a more detailed study of all the witnesses. However, this yet incomplete survey allows some preliminary speculation. Notwithstanding the intentions manifested by Rivers in the Prologue, he decided to invest considerably in Caxton’s press to favour the dissemination of his work outside of the sphere of the court. This policy favoured the acknowledgement of Rivers as a literary and moral authority, which extended also after his death in 1483: Caxton issued a third edition in 1489, followed by Wynkyn De Worde in 1528. This latter edition witnesses the persistence of a literary genre that is traditionally considered “medieval” and that proves to be successful even in one of the most thriving periods of English humanism –​king Henry VIII’s reign. Moreover, the production of multiple manuscript copies after the printed editions indicates that the text was widely read and was considered as valuable as to flank (or integrate) authoritative educational texts such as the Secretum secretorum. Rivers’s prestige deriving from his appointment as tutor of the Prince of Wales elevated his work at the same level of the more celebrated mirrors for princes of those times and made him win great renown, notwithstanding the modern scholars’ neglect. The patronage of Caxton and his technology and the strategy of distribution of the Dicts favoured the success of this work especially among the gentry, who saw this book as a medium to consolidate social relations and gather useful competences for the moral growth of the individual. Owning a copy of the Dicts might mean getting a glimpse of courtly life and education, but also –​and perhaps more importantly –​being able to acquire useful notions for personal edification and the management of one’s own wealth. Thanks to Rivers’s patronage and his policy of dissemination of the Dicts through Caxton’s printing press, also lesser nobility might become as wise as a king.

References Acker, Paul. 1993. “The Craft of Nombrynge in Columbia University, Plimpton MS 259”. Manuscripta 37.1: 71–​83. Bennett, Henry S. 1970. The Pastons and Their England. Studies in an Age of Transition. Cambridge: Cambridge University Press. Blades, William (ed.). 1877. The Dictes and Sayings of the Philosophers: A Facsimile Reproduction of the first Book Printed in England. London: Stock.

56

Omar Khalaf

Blake, Norman F. 1991. William Caxton and English Literary Culture. London: The Hambledon Press. Bühler, Curt. 1948. “New Manuscripts of the Dicts and Sayings of the Philosophers”. Modern Language Notes 63.1: 26–​30. Clanchy, Michael. 1993. From Memory to Written Record. England 1066–​1307. Oxford: Oxford University Press. Cressy, David. 1980. Literacy and the Social Order. Reading and Writing in Tudor and Stuart England. Cambridge: Cambridge University Press. Crotch, Walter J. B. 1928. The Prologues and Epilogues of William Caxton. EETS O.S. 176. Oxford: Oxford University Press. De Ricci, Seymour. 1935–​1940. Census of Medieval and Renaissance Manuscripts in the United States and Canada. With the assistance of W.J. Wilson. Three volumes. New York: Wilson Company. Eckart, Siân –​Robert Rouse (eds.). 2017. The Encyclopaedia of Medieval Literature in Britain. Ames, Iowa: Wyley-​Blackwell. Eder, Robert. 1915. “Tignonvillana inedita”. Romanische Forschungen 33: 851–​1022. Fleming, Peter. 1987. “The Hautes and their ‘circle’. Culture and the English gentry”. In: Williams, Daniel (ed.): 85–​102. Ford, Margaret L. 1999. “Private Ownership of Printed Books”. In: Hellinga, Lotte –​Joseph Trapp (eds.): 205–​228. Hellinga, Lotte –​Joseph Trapp (eds.). 1999. The Cambridge History of the Book in Britain. Volume III 1400–​1557. Cambridge: Cambridge University Press. Hudson, Harriett. 1984. “Middle English Popular Romances: The Manuscript Evidence”. Manuscripta 28.2: 67–​78. Harper-​Bill, Christopher –​Ruth Harvey (eds.). 1992. Medieval Knighthood IV. Woodbridge: Boydell. Hughes, Jonathan. 1992. “Stephen Scrope and the circle of Sir John Fastolf: Moral and Intellectual Outlooks”. In: Harper-​ Bill, Christopher –​Ruth Harvey (eds.): 109–​146. Ives, Samuel A. 1942. “Corrigenda and Addenda to the Descriptions of the Plimpton Manuscripts as Recorded in the De Ricci Census”. Speculum 17.1: 33–​49. Jayne, Sears R. 1956. Library Catalogues of the English Renaissance. Berkeley: University of California Press. Johnston, Michael. 2014. Romance and the Gentry in Late Medieval England. Oxford: Oxford University Press.

The Case of Earl Rivers’s Dicts and Sayings of the Philosophers

57

Keiser, George R. 1999. “Practical Books for the Gentleman”. In: Hellinga, Lotte –​ Joseph Trapp (eds.): 470–​494. Khalaf, Omar. 2011. “Lord Rivers and Oxford, Bodleian Library, MS Bodley 264: a Speculum for the Prince of Wales?”. Journal of the Early Book Society for the Study of Manuscripts and Printing History 14: 239–​250. Khalaf, Omar. 2017. “Dicts and Sayings of the Philosophers”. In: Eckart, Siân –​ Robert Rouse (eds.): 667–​668. Khalaf, Omar. 2018. “The Fragment of Earl Rivers’ Dicts and Sayings of the Philosophers in London, British Library, Add. 60577”. Neophilologus 102.4: 561–​571. Khalaf, Omar. Forthcoming. “The Social Function of a Translation: Earl Rivers, William Caxton, and The Dicts and Sayings of the Philosophers”. In: Rychterová, Pavlina –​Jan Odstrcilík (eds.). Lyte, Henry H. C. M. 1911. A History of Eton College (1404–​1910). London: Macmillan. Mancini, Domenico. 1969. The Usurpation of Richard III. Ed. and trans. C.A.J. Armstrong. Oxford: Oxford University Press. Nichols, John. 1790. A Collection of Ordinances and Regulations for the Government of the Royal Household, Made in Divers Reigns. From King Edward III to King William and Queen Mary. Also Receipts in Ancient Cookery. London: Society of Antiquaries. Oates, J.C.T. 1960. “English Bokes Concernyng to James Morice”. Transactions of the Cambridge Bibliographical Society 3.2: 124–​132. Orme, Nicholas. 1984. From Childhood to Chivalry. The Education of the English Kings and Aristocracy 1066–​1530. London –​New York: Methuen. Orme, Nicholas. 2005. “Education and Recreation”. In: Radulescu, Raluca –​ Alison Truelove (eds.): 63–​83. Parkes, Malcolm B. 1991. “The Literacy of the Laity”. In: Parkes, Malcolm B. (ed.): 291–​297. Parkes, Malcolm B. (ed.). 1991. Scribes, Scripts and Readers. Studies in Communication, Dissemination and Presentation of Medieval Texts. London: Hambledon. Pollnitz, Aysha. 2015. Princely Education in Early Modern Britain. Cambridge: Cambridge University Press. Putter, Ad. 2000. “Introduction”. In: Putter, Ad –​Jane Gilbert (eds.): 1–​38. Putter, Ad –​Jane Gilbert (eds.). 2000. The Spirit of Medieval English Popular Romance. Harlow: Longman.

58

Omar Khalaf

Radulescu, Raluca. 2005. “Literature”. In: Radulescu, Raluca –​Alison Truelove (eds.): 100–​118. Radulescu, Raluca –​Alison Truelove (eds.). 2005. Gentry Culture in Late Medieval England. Manchester: Manchester University Press. Rees Jones, Sarah –​Richard Marks –​Alistair J. Minnis (eds.). 2000. Courts and Regions in Medieval Europe. Woodbridge: Boydell. Richmond, Colin. 2000. “The Pastons and London”. In: Rees Jones, Sarah –​ Richard Marks –​Alistair J. Minnis (eds.): 221–​226. Riddy, Felicity. 1987. Sir Thomas Malory. Leiden: Brill. Rychterová, Pavlina –​Jan Odstrcilík (eds.). Forthcoming. Medieval Translator. Medieval Translations and their Readerships. Proceedings from the Cardiff Conference on the Theory and Practice of Translation in the Middle Ages. Turnhout: Brepols. Salter, Elisabeth. 2012. Popular Reading in English. C. 1400–​1600. Manchester: Manchester University Press. Sutton, John W. 2006. The Dicts and Sayings of the Philosophers. (Teams Middle English Series). Kalamazoo, Michigan: Medieval Institute Publications. Trapp, Joseph. B. 1999. “Literacy, Books and Readers”. In: Hellinga, Lotte –​ Joseph Trapp (eds.): 31–​43. Truelove, Alison. 2005. “Literacy”. In: Radulescu, Raluca –​Alison Truelove (eds.): 84–​99. Weinberg, S. Carole. 2005. “Caxton, Anthony Woodville, and the Prologue to the Morte Darthur”. Studies in Philology 102.1: 45–​65. Williams, Daniel (ed.). 1987. England in the Fifteenth Century. Proceedings of the 1986 Harlaxton Symposium. Woodbridge: Boydell. Wilson, Edward –​Iain Fenlon (eds.). 1981. The Winchester Anthology: A Facsimile of British Library Additional Manuscript 60577. Cambridge: D.S. Brewer. Youngs, Deborah. 2005. “Cultural Networks”. In: Radulescu, Raluca –​Alison Truelove (eds.): 119–​133.

Patrizia Lendinara

On the Trail of Bibbesworth Abstract: The Tretiz of Walter of Bibbesworth exerted quite an influence on literature over the two centuries that followed its composition. It served as the basis for two long works in French –​accompanied by a full translation in English –​which were employed as learning tools, while its effect was evidently felt in a number of bilingual (or trilingual) glossaries, such as the Glossary in Cambridge, St John’s College, E.17 (120). Later French language learning tools circulating well into the fifteenth century also display a working knowledge of the Tretiz. Keywords: Walter of Bibbesworth, Tretiz, Glossaries, Loanwords, Homonyms.

1. A Work that Left Its Mark The Tretiz ‘pur aprise de langage’ written by Walter of Bibbesworth enjoyed a wide popularity in England. Not only is the Tretiz attested by a remarkable number of manuscripts, but its influence is evident in a number of French language learning texts dating from the fourteenth and fifteenth centuries. The overlapping with the Nominale in Cambridge, University Library, Ee. 4.20 is considerable and the bulk of the anonymous treatise Femina is taken from the Tretiz. A number of glossaries pick up the French words belonging to the lexical fields covered by the Tretiz (e.g. names of animals and their cries, trees, food and farm implements) and their ME renderings provided by the glosses which accompany the work in several manuscripts. In the Manières de Langage, whilst dealing with the vocabulary necessary for English travellers in France, a few sections follow verbatim (albeit in abbreviated form) the similar sections of Bibbesworth’s work. The Ortographia Gallica (Johnston 1987) employs the lists of French words cramming the lines of the poem, focusing on the homonyms which are subtly deployed by Bibbesworth. All these pieces of evidence bear witness to the currency and influence of the Tretiz in England during the whole of the fourteenth century and beyond.

60

Patrizia Lendinara

2. The Tretiz On the basis of the other works composed by Walter of Bibbesworth and what is known of the author’s life, the Tretiz ‘pur aprise de langage’1 was written in 1270 at the latest.2 The Tretiz is addressed to people who had some knowledge of French and, in its Prologue, is presented as a tool for the children of Dionyse (Dionysia or Dionisie) de Munchensy, who apparently wished to teach French to her two stepchildren, Joan and John, and to her own son, William; thereby likely improving her own French in the process. The lines of the long composition relate to a number of areas of French vocabulary: parts of the body, clothing, names of animals (as well as collective names for animals and their voices, with particular regard for birds), flowers, fruits, trees, agricultural pursuits, food and table settings (Jambeck 2005). The author’s command of French lexis is remarkable, and he makes skilful use of French homonyms and near-​homonyms, several sets of which are employed in his work. For example, the lines bring together a play-​on-​words such as arable, which means ‘maple’ as noun, and ‘arable’ as adjective. The same applies to near-​ homonyms such as greve ‘parting in the hair’ and grive ‘fieldfare’. Such word-​play seems to imply an audience which already had a good understanding of French.3 The Tretiz undoubtedly vouches for the continued use of French in aristocratic households, where the French language was retained along with English lands for generations following 1066. The Tretiz survives partially or wholly in seventeen manuscripts, which were produced between the mid-​thirteenth and the early fifteenth centuries. These codices were copied in separated areas of England (Koch 1934: 58). The length of the work varies from one manuscript to another, as some lines were omitted or substituted and whole sections reshaped or copied in a different order. The very nature of the work was such as to invite additions and omissions by copyists; 1 “Tretiz […] pur apprise de langage” is how the work is defined in the prologue of Cambridge, University Library, Gg. 1.1 (G) (sec. xiv1/​4). 2 Baugh (1959) dated the composition to 1240–​1250. William Rothwell (1982: 282) endorsed the dating; he, however, would push back the assumed date of composition over time in his following essays and in his edition. Owen (1929: 41) would date the work to 1280–​1290. 3 The lack of a comprehensive editorial treatment hinders the comprehension of the Tretiz, that will be properly assessed only when all the surviving manuscripts of Bibbesworth and their Middle English glosses will be published. The project by Dr. Thomas Hinton of the Exeter University moves in this direction. The Tretiz is attested to by seventeen manuscripts, a part of which are fragmentary.

On the Trail of Bibbesworth

61

the different versions display a certain independence. The longest version of the Tretiz amounts to about 1200 lines. Walter’s first text, the one presented to Dionyse de Munchensy, was likely quite close to Cambridge, Trinity College, O.2.21 (T) (s. xiv), which is prefaced by a dedicatory letter. In the recent research on the Tretiz, Dionyse has gradually faded into the background, while due attention is paid to the environment where the manuscripts were produced in addition to the contents of the codices themselves. For several scholars, the Tretiz is a key text for understanding the status of French in England in the thirteenth century. According to Thomas Hinton’s recent re-​interpretation of the work, the Tretiz reflects “a growing internationally-​ shared confidence” in French “as a language of culture” (Hinton 2017: 863). Hinton takes a determined stand against the view of the Tretiz as an essentially functional text primarily concerned with language acquisition.4 Notwithstanding the original aim of the Tretiz, the work provided a classified exposition of French vocabulary (Rothwell 1982: 282), joined together in complete sentences and in versified form (octosyllables), providing a rich store of (partly bilingual) lexicographical material which would not escape the attention of later generations of masters.

2.1. The glosses to the Tretiz One of the remarkable features of the majority of the manuscripts of the Tretiz is the presence of a high number of English glosses. The amount of glossing varies from one manuscript to another. One of the codices, London, British Library, Sloane 513 (s. xv1/​4), does not bear any gloss. That of the Tretiz is a peculiar apparatus of glosses, whose presence and role has received differing assessments by the scholars. According to a school of thought represented (amongst others) by Koch (1934), the glosses stem from the author himself; on the other hand, the most extreme appraisal given to these glosses is that of Ingham, who in 2014 stated that there is no evidence that the glossing was the work of Bibbesworth. As for Ingham’s viewpoint (2014: 435), the original was not accompanied by glosses. This statement squares with Ingham’s general opinion about the Tretiz, according to which “the poem in its original 4 Hinton also refutes the older interpretation that the Tretiz was meant to teach the lexicon of French that would be useful to a once grown-​up adult for running his family estate and was hence focused on agricultural or household vocabulary (Hinton 2017: 864–​865).

62

Patrizia Lendinara

form formed no part of a syllabus of instruction via English” (Ingham 2014: 453 and 457). In the representation suggested by this scholar, the glosses, as we now have them in the manuscripts of the Tretiz, were written at a later date and stem from “unsystematic private study of self-​taught learners, written cumulatively in the margins of the surviving manuscripts at a later date” (Ingham 2014: 438). The presence of the glosses as an integral part of the work is, on the other hand, justified in the last words of the prologue to the Tretiz, where it is written “Dounc tut dis troverez vous primes le fraunceis e puis le engleise amount” (You will see the French first, and the English just above it) (Rothwell 2009: 1).5 In a number of cases, the English glosses were copied onto specially ruled columns and were accommodated on the right-​most edge of the lines. There are also interlinear glosses, and some codices have both interlinear and right-​most edge glosses (Knox 2013: 352). In my opinion, the core of the English glosses was part of the author’s original design, and the majority of the glosses were transmitted along with the text of the Tretiz. Although the apparatus of glosses –​remarkable for its continuity for about two centuries –​is not homogeneous, a good number of English glosses occur unchanged in most of the manuscripts. Some codices are more heavily glossed, such as Oxford, All Souls College 182 (O), and some particularly idiosyncratic glosses may stem from later scribes. These scribes were also responsible for the updating of English glosses, the addition of either the determinate or the indeterminate article before the gloss, and also the replacement of one English word by another with the same or a similar meaning. Of course, they were also responsible for the omission and addition of glosses. What needs to be reassessed is the reason why Bibbesworth accompanied the lines of his poem with a number of English glosses. Far from being a mere didactic tool to help anglophone readers understand French, the Tretiz proves to be a witty exercise, noteworthy for its jocular vein, which permeates the whole literary output of Bibbesworth as it is emerging in the most recent studies.

5 Unless otherwise specified, all references are to the G version in Rothwell (2009), which is 1140 lines long. The prologue is found in five manuscripts (A, C, G, 8, and Y). The letters used to refer to the manuscripts are those adopted in Owen’s edition of the Tretiz (which takes into account fourteen codices); Koch uses a different set of symbols. Y stands for Yale, Beinecke Library, Osborn a56 (s. xiv1/​2) and Bo for Oxford, Bodleian Library, Bodley 39 (s. xiv1/​2).

On the Trail of Bibbesworth

63

3. The Cambridge Nominale and Femina Not only is the Tretiz attested by a remarkable number of manuscripts, but its popularity and influence is evident in a number of works dating from the fourteenth and fifteenth centuries. Beforehand a distinction should be drawn between the Cambridge Nominale and Femina and a number of other glossaries and works composed in England during these two centuries. The relationship between the two former works and the Tretiz is evident and was pointed out since the first edition of the two compilations. The influence of Bibbesworth on other Nominalia is not so pervasive, but still relevant as the instances discussed below will demonstrate. A Nominale, which derives in large measure from the Tretiz, is only preserved in Cambridge, UL, Ee. 4.20, ff. 162r–​164v.6 The manuscript that contains such work is a miscellany compiled at St Albans, ca. 1380–​1400, featuring both items of legal and administrative nature and texts employed for teaching French. It encompasses works in Latin such as the Orthographia Gallica, as well as a list of colours and numbers in French.7 The French Nominale is versified and its couplets have a quite repetitive structure (888 lines long in Skeat’s edition). The English equivalent in prose is set to the right, line by line, and proves that this adaptation of Bibbesworth’s Tretiz was explicitly meant to teach French by using English. The Nominale treats several topics in an order other than the Tretiz and –​as pointed out by Evans (1993: 111) –​greatly expands some lexical fields; for example, that of bird names. It was likely part of a wider course of study and was copied alongside works that date to the second half of the fourteenth century. A short time-​span between the composition of the Cambridge Nominale and the Tretiz was surmised by Skeat, who would date the manuscript to 1340, but considered the Nominale a copy of an original dating to the thirteenth century. Skeat, who declared his interest in the ME lexicon of the Nominale, chose to number the lines so as to distinguish the French and the English part, a choice which misrepresented the original lay-​out. In the original, French and English are written in the same line but with a break between each other, Corps teste et hanapel Body henede et henepanne Et pal gesceant sur la peal And here growede onge skyn. (Skeat 1903–​1906: ll. 1–​4) ‘Body, head and brainpan, /​and hairs growing on the skin’

6 Ff. 133r-​135v, as to the double pagination of the codex, see Skeat (1903–​1906). 7 See Dearnley (2016: 153–​154), for the manuscript context. Dearnley relates the contents of the codex to the school of Thomas Sampson in Oxford (a school of “Business French”, operating independently of the university).

64

Patrizia Lendinara

This Nominale features several rare words which will appear again both in Femina and the following glossaries and which have no counterpart in the Tretiz; for example, ME stokdouve which renders A-​N columbe ramer ‘wood-​pigeon’ (ll. 803–​804) or ME turtildouve which glosses A-​N turtre ‘turtle-​dove’ (ll. 805–​806). The latter reworking of the Tretiz goes under the title of Femina Nova or simply Femina, and is preserved in a single manuscript, Cambridge, TC, B 14.39/​ 40, dating from the early fifteenth century (Dearnley 2016: 155). In Femina, each couplet in French is followed by a full translation into English. This is how the lines appear in the manuscript: the English component is not a gloss but a translation of the French couplet, and the two versions are alternately copied, Beau enfant pur apprendre En franceis devez bien entendre ffayre chyld for to lerne In frensh 3e schal wel understand (Rothwell 2005: 1). ‘Fair child, to get educated,/​you should master French well’

Femina draws fully from the Tretiz of Bibbesworth but employs its material to a different end while addressing a different audience. It has been estimated that about eighty per cent of the text of the Tretiz found its way into Femina (over 800 verses as against ca. 1200).

4. The Glossaries A glossary whose dependence from the Tretiz is at first sight evident is that in Cambridge, St John’s College, E.17 (120) (cf. Acker 1993 and Lendinara 2015). This short glossary with French lemmata and English interpretamenta is probably an excerpt from a larger topical compilation. The glossary, which is copied in the upper part of f. 126r, is the only piece of the manuscript not written in Latin; the short compilation is indeed one of the few French-​English glossaries from medieval Britain. The part of the codex containing the glossary (ff. 211–​ 216) has been dated to the fourteenth century.8 The majority of the French lemmata of this glossary (as well as their English renderings) have a counterpart in the Tretiz of Bibbesworth and a cluster of texts depending on it. Of the 38 entries of the glossary, 23 have a parallel in the Tretiz.

8 For the manuscript, see James (1913: 153–​155) and the St John’s College website: http://​ www.joh.cam.ac.uk/​library/​special_​collections/​manuscripts/​medieval_​manuscripts/​ medman/​E_​17.htm. As it now stands, the codex is composed of seven or eight libelli or fragments put together at an uncertain date. See Thomson (2013: 166–​168).

On the Trail of Bibbesworth

65

In 11 instances, the overlapping concerns both the French lemma and its gloss (nos. 4, 5, 9, 12, 13, 18, 19, 22, 25, 31, and 36); in a few cases, the ME gloss occurs only in part of the codices of the Tretiz or in additional or substitute lines (nos. 1, 2, 6, 10, 21, and 24, on which see below). These cases would indeed be the most interesting to investigate when all the versions of the Tretiz become fully available, which will hopefully happen in the next few years. In two instances, the Tretiz and the glossary feature the same French lemma, but a different ME gloss (nos. 8, 20), or vice versa (nos. 7, 15, 30, and 38). In one of these last instances the ME gloss, wōdewāle (no. 30), occurs in both the Tretiz and in the glossary under consideration; however, in the St John’s glossary, the original French lemma has been replaced, likely owing to a scribal error, by a ME one: heghwal ‘woodpecker’ (“Eywal . Wodewale.”).9 Three entries of the glossary (nos. 16, 26, and 29) have a counterpart in Cambridge Nominale and repeat both the French lemma and its ME translation. With regards to the dating of the glossary, the close relationship with the Tretiz puts its date to at least the second half of the thirteenth century, if not later. The French loanwords aubē̆ltrē ‘white poplar’ and cokkou ‘cuckoo’ also argue in favour of such dating. As far as the localization is concerned, a few of the glossary’s items are recorded exclusively in documents from the northern counties of England. A few hints would point to Yorkshire.10 Undoubtedly connected to the aforementioned texts is the Nominale in Cambridge, UL, Add. 8870, ff. 30r-​32v (s. xv) (Baker 1989). This long topical glossary has rubrics in Latin; the lemmata are in French and the interpretamenta are mostly in Latin. In a number of cases, the interpretamentum is given in English instead, whether preceded by the label anglice or not. A remarkable case is that of “une engle[n]‌tier que porte lez pipe rouges” (no. 391), which is also left without a gloss and is a verbatim quotation from line 513 G of the Tretiz. The detailed list of the parts of the body that opens this glossary repeats a large number of the words used by Bibbesworth in the order in which these words are listed in the Tretiz; the glosses are also the same. However, the glossary has also other sources, from which it draws, for example, words referring to the body of women, and pudenda in general. The relationship between the Add. 8870 Nominale and the glossary

9 MED H, pp. 576–​577: hegh-​wal; see Kitson (1997: 496–​497), and Lockwood (1984: 82). 10 Laing (2003: 33) did not advance any proposal as to the dialectal provenance of the glossary, suggesting a dating to “?a. 1300”. The glossary has been dated to ca. 1275 by the AND. The bilingual word list as it now stands in St John’s College 17 is a copy of a former (and likely larger) glossary, which might predate it by a century.

66

Patrizia Lendinara

of St John’s College is very close and the two have seventeen entries in common (Lendinara 2015: 124–​126). In the trilingual glossary (Latin-​French-​English) in London, Westminster Abbey 34/​11 (s. xiv) there occur several entries belonging to this cluster of works (Hunt 1990). This compilation shares fifteen items with the glossary of St John’s College. A fragmentary topical glossary, it lists a series of iron tools and utensils (for cooking, etc.), followed by other tools (as to the rubric ‘de lingno (sic) et de calibo et cipreo’) and a long section on bird names (‘Nomina avium’); the glossary breaks up after a few names of wild animals, which belong to a discrete section. This section opens with birds of prey, both nocturnal and diurnal (which were not included in the Tretiz), followed by large domestic fowls and game birds: several entries have a counterpart in ll. 711–​802 (devoted to the woodland birds) and elsewhere (e.g. “altio: un assy, a wodecok”: aciez woddekockes line 1129 G; ascye wodecock line 551 T; “grus: un grue, a grane (l. crane)”: grues cranes ll. 211 –​without gloss –​, 223, 250, 1119 G, 188, 828 T). Two other entries are worth highlighting. The first is: “auca josaria (l. ro­saria): owe joer (l. roser), a wilde gos”, which features both the unusual Latin lemma and the French rendering shared by the Tretiz-​cluster of texts.11 The entry “aucerulus: un osilou, a goseling” is similarly exclusive. The Westminster glossary also features the rare columbe ramer: “palumbus: columbe ramer, a stok dowe”. Both glossaries might either descend from the (class-​)glossary used by Bibbesworth (and the Cambridge Nominale) or directly from one of these works.

5. Variations in Glosses The apparatus of glosses which accompany the Tretiz in its various manuscripts is not homogeneous as regards the rendering of some words of the poem. The

11 See Tretiz line 787 G, 829 T; Cambridge Nom. ll. 795–​796; Femina (Rothwell 2005: 60); Westminster gloss. “auca josaria: owe joer, a wilde gos”; Add. 8870 Nom. no. 326 “un ouue roser, anglice a wylde gose”; St John’s gloss. no. 20: “Ouwe roser . Redlag.”. Latin rosera refers to a species of goose; Old French roziere ‘a wild sort of goose’ occurs, amongst others, in the translations of the treatise by Frederic II (s. xiiiex), while roser ‘wild’ (of geese) is used in the translation of the hawking treatise by Ghatrif. It also occurs in Letter no. 341, which gives a long list of bird names (Legge 1941: 406); and in the Geste de Blancheflour et de Florence line 82 (roseer): for this poem and the Tretiz, see below. A derivation from Old French ros ‘reed’ (FEW 16, 681b) is much more likely than that from Latin rosa ‘rose’ (with allusion to the colour of the feathers) (FEW 10, 481; FEW 21, 244a is to cross out).

On the Trail of Bibbesworth

67

manuscripts of the Tretiz originate in different areas of the country and, in some instances, the ME glosses might reproduce a dialectal differentiation.

5.1. Snipe/​Snite The difference in glossing has its follow-​up in the entries of the glossaries which draw on the Tretiz. For example, in the St John’s glossary, the English gloss for ‘snipe’, a small wading bird, is snipe (“Bekaz . Snipe .”: no. 24),12 which also occurs in the Add. 8870 Nominale: “un becaz, anglice snype” (no. 322), whereas, in the Westminster glossary, there occurs the variant snite: “ficedula: un bekaz; a snyte”. ME snīte, which is now limited to Devon and Cornwall,13 continues OE snīte (also wūdusnīte);14 on the other hand, snīpe is commonly identified as a loanword from ON -​snípa as in mýrisnípa ‘moor snipe’ (Cleasby –​Vigfusson 1957: 441 and 575). The by-​name of this long-​billed marsh bird, which in due course prevailed over its synonym, might otherwise be native in English. The alternative use of snīte and snīpe reproduces the variety of the glosses to this line of the Tretiz, E li oisel ke ad noun bekas (snyte) (l. 795) ‘And the bird whose name is snipe’

T (line 593), A, and P (snip) have snipe, whereas G (line 795), B, 8, C, Y, and O have snīte. The word bekas is left unglossed in Bo. The Cambridge Nominale reshapes the line at the end of the section on birds: “Bekas et musenge Snyte and titemose” (snipes and tits) (ll. 819–​820), whereas Femina repeats the Tretiz, but, in the translation, employs a loanword from French, bēke (Rothwell 2005: 61).15

12 MED S, pp. 58–​59 snīpe [Prob. ON: cp. OI mӯri-​snīpa, Norw. myrsnipa, strandsnipa]; snīte; see Lockwood (1984: 103 and 142–​143); Kitson (1997: 500); Lendinara (2015: 113, notes 31–​35, 128–​129). According to Björkman (1900–​1902, II: 255), snīpe is probably native in English; Kries (2003: 213 and 220–​221) gives the term (Middle Scots myresnipe) as a loanword. 13 Cf. snipe (Wright 1898–​1905, V: 584) and snite (Wright 1898–​1905, V: 586). 14 The gloss in London, BL, Stowe 57 (s. xii2, Peterborough) –​whose content, as far as ff. 156v–​165r are concerned, awaits a proper description –​bears no evidence as regards to this bird name, because only the first three letters of the word are available: “sni[…]” (Hunt 1991, I: 23). 15 bēke (also bebibek) is a loanword from French bekaz. It has been remarked that, differently from the Tretiz, Femina employs a large number of ME words of French origin. For its part, the Cambridge Nominale uses native English words to a far greater extent than Femina.

68

Patrizia Lendinara

5.2. The Blackbird As regards another bird name, the majority of manuscripts of the Tretiz use the traditional name for the blackbird, osel, to gloss A-​N merel,16 En branche set le menue merle. (osel) (line 717) ‘The little blackbird sits on a branch.’17

ME ōsel is also the rendering provided by the Cambridge Nominale and the Westminster glossary: “merulus: un mercel, an osil”, whereas the Add. 8870 Nominale leaves this bird name unglossed: “un merle, –​” (no. 313). The St John’s glossary, in this case, employs a different ME word: “Merel . Suswart.” (no. 21). Again, sūswart18 might stem from the scribe of St John’s College 120, who substituted the original interpretation, possibly employing a more familiar local term. The bird was also known by other names in Middle English: throstel –​ which is used in Femina –​, throstelcok, blakbrid,19 and merule (a loanword from Latin merula), but sūswart remains isolated. However, a manuscript of the Tretiz, Cambridge, Corpus Christi College 450, features the same gloss as the St John’s glossary, sousuart.20 The codex is a miscellany of treatises, poems, historical documents and legal records, written in Durham between the years 1290 and 1320. It contains the shorter version of the Tretiz, which belongs to the family of T (Cheney 1972). Sūswart can be explained as Norse-​derived, as is the case with other ME words of the St John’s glossary (Lendinara 2015: 106–​115). This word has a striking counterpart in súsvǫrt ‘nightingale’ which occurs in the Old Norwegian translation of the Saga of Barlaam and Josaphat.21 16 MED O, pp. 301–​302: ōsel(e. The compound ōselbrid is also recorded; this term goes back to OE ōsle ‘blackbird’. 17 þe oosel brid T (line 543), an hosel brit A, usel 4, the lutel oslebrid C, luitel osel brid Y; many osell O (O has also another occurrence of the word before line 1131: merles: oisels); Cambridge Nominale ll. 799–​800: merle /​osele (the gloss blacbrid was added in the manuscript); Femina (Rothwell 2005: 54–​55): merele: þe þrosshe. 18 MED S. 18, p. 1153: sūswart [ON: cp. OI sūsvort]. 19 As to Lockwood (1984: 30), blackbird is first mentioned in 1486, but see the additional gloss to the Cambridge Nominale. 20 See https://​parker.stanford.edu/​parker/​catalog/​bc961dv0378. Note, however, that CCCC 450 and the glossary in St John’s College do not share any other exclusive match, and many of the words occurring in the glossary do not bear any gloss in this manuscript of the Tretiz. 21 “einn veiði maðr. tok einn fugll með list. þann er heitir filomena a latinv. en a norreno heitir susvort. sumir kalla oc niktigalo” (Rindal 1981: 30, ll. 22–​23). Copenhagen, Den

On the Trail of Bibbesworth

69

5.3. A Brake. Feryn In another instance, a different gloss is offered in conjunction with the more frequent one. This is the case of the gloss to feugere ‘fern’, Eschauffez le fourn de feugere (feron) Pur defaute de littere. (ll. 396–​397) ‘Heat up the oven with ferns /​if there is no straw.’

Whereas, in the majority of the manuscripts, A-​N feugere is glossed by ME fē̆rn and variants (ferne, feryn, vern, veern), two codices provide an alternative rendering and use ME brāke(n).22 London, BL, Arundel 220 (A) (s. xiv1/​4, East Anglia) has two glosses for feugere: a brake. feryn. In Yale, Beinecke Library, Osborn a56 (Y) (s. xiv1/​2), in the margin of the line containing feugere, there is written feern ou brucve (f. 11r). The latter word is underlined; above there is another word enclosed by a square, which, unfortunately, has been cut by the trimming of the folio (be[…]). There are a few other instances in which the conjunction ou ‘or’ is used between two ME glosses to the Tretiz in Y. A and Y belong to different branches of the tradition of Bibbesworth’s work; but we will be in a position to unravel the relationship between its codices and their glosses only when we have a good grasp of the entire manuscript tradition. ME brāke(n) ‘coarse fern’ has been related to Swed. bräken and Dan. bregne, both meaning ‘fern’,23 and classified as a northern England word (Björkman 1900–​1902, II: 231–​232; Kries 2003: 227). However, ME brāke(n) might also be brought back to OE *bracu ‘fern’, which is recorded in the compound OE fernbracu ‘fern’ and OE *bræcen (derived from *bracu ‘fern’ with the suffix -​ en), possibly evidenced in place-​names such as Bracken in the East Riding of Yorkshire (Hough 2003: 48–​49 and 67–​68). No ancient Norse words are attested, but it is possible that an ON *brakni was in use in the northern counties of England and might have contributed to reinforce the already present OE *bracu.

Arnamagnæanske Samling, AM 230 fol., f. 23r (s. xiv2) has the variant reading svsvavrtt; see Ordbog over det norrøne prosasprog, http://​onp.ku.dk, s.v. súsvǫrt and Cleasby –​ Vigfusson 1957: 605. The passage in the Barlaams ok Josaphats saga stands alone in the old literature, where the word is not otherwise recorded; súsvǫrt is preserved in provincial Norse sysvorta (sisvorta, svisvorta) ‘turdus torquatus’. 22 fernT (line 312), vern P, veern C, ferne O; in Bo, P and 8 the word does not bear a gloss, in 4 the folio is cut; Femina: spayes (Rothwell 2005: 32); St John’s gloss. no. 10: “Fugere. Brakan .”. 2 3 MED B, p. 1109: bracken, brāke(n [ON; cp. Swed. bräken, Da. bregne.]; OED, s.v. bracken.

70

Patrizia Lendinara

As to brāke(n) ‘fern’, the word is attested to since the thirteenth century,24 side by side with its synonym fē̆rn, which continues OE fearn, for the entire ME period, with a preponderance of occurrences in the N. W. Midlands (e.g. Cleanness [London, BL, Cotton Nero A.x], line 1675) and in Norfolk.25 The early English dictionaries generally provide both synonyms, either duplicating their entries or by a cross-​reference, while the Promptorium parvulorum, whose compiler explicitly states that he uses the dialect of Norfolk for the English headwords, has only “Brake, herbe: ffilix, -​cis” (Mayhew 1908 [1973]: 46).

5.4. The Case of Fila(u)ndre One of the features of the Tretiz is its exploitation of homonyms to a didactic end, often with light and humorous overtones. Glosses have a relevant role in decoding the sets of homonyms, but mistakes might easily intrude upon the process of copying. A further and still unrecognized homonym might have been used by Bibbesworth. A-​N fila(u)ndre is, on the one hand, a term designating some sort of thread or web, usually identified as the gossamer, and, on the other, the name of a bird, the starling. In the former occurrence (line 124), as to the context, filaundre refers to an impalpable web and was chosen by Bibbesworth to set up a witty comparison with the ginger seeds. A large quantity of the worthless filaundre is nothing compared to a small amount of the precious ginger, Kar meuz vodroie petite poiné De gengevere bien trié Ki ne ferroie cent galeins (þespen) De filaundre tut pleins. (ll. 121–​124) ‘Because a small handful /​of well-​chosen ginger is worth more /​than a hundred double-​ handfuls /​full of gossamer.’

Bibbesworth might have chosen a spice as a symbol of the good life and the lines might conceal a mocking intent. Note that the comparison with the ginger seeds

24 The gloss wylde brake was added (in a hand of the thirteenth century) to the drawing of a polypodium in the Herbarium of Pseudo-​Apuleius in Oxford, Bodleian Library, Bodley 130, f. 24v (s. xi2, Bury St Edmunds, Suffolk). Another early occurrence is provided by a gloss to the Unum omnium in Oxford, BodL, Rawlinson G. 96 (s. xiii): “filix feuger, brake” (Hunt 1979: 167). For further occurrences, see Hunt (1991, II: 15, 17, 19, 20, and 108); Nominale in London, BL, Royal 17.C.XVII: “Hic felix, -​cis ae brakyn” (WW 644, 17); Mayer Nominale: “Hec felix, -​icis, media corrupta, brakyne”; “Hoc felicetum, Hoc filacerium, a brakynbuske” (WW 712, 1–​4). 25 See the examples in MED B, p. 1109.

On the Trail of Bibbesworth

71

is not deployed by the author of the Cambridge Nominale, where there occurs instead a line on the mermaid and her tail. Ginger was highly prized during the Middle Ages, when spices were luxuries. Following the development of commerce of the Hanseatic League, spices (which were carried from Constantinople on the Russian rivers and, by ways of Kiev and Nizhni Novgorod, to the Baltic Sea) reached England. The Baltic routes would trade spices such as ginger in England in exchange for wool. The major spices were pepper, ginger and cinnamon: they were undoubtedly expensive and their cost might be understood in terms of comparison with prices for other goods. By the late Middle Ages, a pound of ginger costed as much as a sheep (according to period price tables).26 It is this standard unit of measurement of the value of ginger, that is ‘a sheep’, which suggests that by filaundre, Bibbesworth did not originally mean the gossamer, but either a mass of wool (or waste of wool) or a lump of whitish threads. As a matter of fact, the MED dates the meaning of ME gossomer as ‘fine filmy substance, consisting of cobwebs’ to the Tretiz, while the sense ‘as a type of something light, trivial, or worthless’ is ascribed to Lydgate (see below). French filandre (FEW III, 535; TLF VIII, 877) is attested to with the meaning of ‘ficelle’ at the end of the thirteenth century and, a century later, of ‘threadlike intestinal worms that afflicted hawks’. Hence the sense currently suggested for both A-​N fila(u)ndre and ME gossomer as these words are used in the Tretiz27 needs some refocusing. French filandre is a derivative of filer ‘to spin’ (FEW III, 535b, 537b: filum). In the late Middle Ages, the term is attested in the meaning ‘decorative threadlike object’, ‘sort of fishing dam’, and ‘intestinal worm that afflicted hawks’. The first occurrence of filandre as ‘fil léger sécrété par certaines araignées’ or ‘fil vaporeux qui flotte dans l’air (→ fil de la vierge)’ is tracked down to the Tretiz of Bibbesworth by several dictionaries, but this rather looks like a backdating. As to the AND, it records three different filandre: for both filandre ‘gossomer’ and filandre ‘starling’ (see below), the earliest reference is to the Tretiz. The AND also registers a filandre ‘cudweed’ on the basis of a medical receipt, “filago: filandre, chaunette, horewort” (Hunt 1993: 184).

26 Munro, quoted by Pilcher (2012: 337), estimated that, in London, in 1200, it took nearly 8.6 days’ pay of a master craftsman to buy a pound of ginger. 27 Filaundre (line 124) is left unglossed in G and P; as to the other manuscripts, they feature these glosses: fol of gosesome(r) T, gosesomer A, of gosesomer C, gossamer Y, gossomer O, gossesmner B; of gossamer R, gosesomer 8, gossummer Bo.

72

Patrizia Lendinara

Another occurrence of filaundre (which reminds us of the above-​mentioned receipt) was signalled by Thurot (1868: 529), who printed a number of glosses from Paris, Bibliothèque Nationale de France, lat. 14745, f. 73r (s. xiii, abbey of Saint-​Victor in Paris). The codex, which contains, amongst others, the Doctrinale puerorum of Alexander of Villedieu with commentary (ff. 47r–​84r), was likely copied by an English hand, and bears glosses in English and French written in columns on the right margin of the text. A gloss to thyma in line 1856 of the Doctrinale reads: “Hec tima erba est cuius flos dicitur epitimum. Hic timus in singulari genus est floris, secundum quosdam genus est piscis, et in plurali hec tima dicitur. Filaundre”.28 Epithyme is the name of small shrubs and herbs, which are parasitic to thyme; it is an annual rootless twining herb. In another section of the Tretiz, Bibbesworth again employs the word fila(u) ndre within the description of the woods and its birds. The term is chosen with the playful intent of combining and rhyming three words in -​dre: filaundre chalaundre, and salemaundre.29 The sequence of the three terms is somehow awkward,30 but the second is a songbird of the lark species31 and the first also might be a bird. A few manuscripts of the Tretiz gloss filaundre in line 719 by ME star(e) ‘starling’;32 other manuscripts, however, repeat the gloss employed in the earlier passage, ME gossamer,33 28 For a similar interpretation, which stems from Dioscorides, De materia medica, CLXXIII De Epitimyo, see Isidore, Etym. XVII.ix.13: “Epithymum Graecum nomen, quod Latine dicitur flos thymi; nam flos Graece θύµον vocatur: est autem flos thymbrae similis”, followed by Hugutio, Derivationes, E 89 (Cecchini et al. 2004, II: 384): “(7) et cum thymum quod est flos grece, et dicitur epythimum, idest flos thymi, et est thymbre similis et dicitur sic quia soleat superponi a medicis aliis adiutoriis precedentibus; unde epythimicus, -​a -​um, idest”; and Sinonoma Bartholomei (Mowat 1882: 19b): “Epithumi, fila quaedam sunt quae nascuntur super thimum. Nascuntur etiam super alias herbas et frutices sed eo magis utimur quod super thimum. Calidum est et siccum in secundo gradu”. Alphita (Mowat 1882: 57b) has a similar interpretation. 29 Salamanders were usually associated with fire in antiquity and the Middle Ages, e.g. in the Bestiary. 30 Evans (1993:110, note 8) would justify the presence of filandre in line 717 on the basis of the commonplace connection of birds and insects in the Middle Ages, and seems to suggest that the word refers to the ballooning or sailing spiders borne aloft by winds on gossamer filaments. Also the AND adds the tag [insects] beside the meaning ‘gossamer’ (but the reference is to line 124). 31 A Mediterranean species renowned for its song. In Middle English glosses, A-​N chalaundre is rendered by wodelarke ‘woodlark’. 32 stare G (line 719), star B, star C, stare 4, þe stare Y. 33 gosesomer T (line 547), gosesomer A, gossamer P, gossomer O, Bo has no gloss.

On the Trail of Bibbesworth

73

Uncore il ad ausi filaundre (stare) E le oysel ki ad noun chalaundre. (wodelarke) Au four meint le salemaundre. (criket) (ll. 719–​721) ‘And there is also a starling /​and the bird that is called woodlark. /​The salamander is in the oven.’

ME star is quite common in Middle English with both the meaning ‘starling’ and ‘thrush’ (also OE stær glossed both sturnus ‘starling’ and turdus ‘thrush’). The choice of the gloss stare in some of the Bibbesworth’s manuscripts is the starting point for the inclusion of the French word (and/​or the corresponding Latin word) interpreted as stare in a few glossarial compilations and works drawing on the Tretiz.34 The Cambridge Nominale, in this case, follows the Tretiz in the line: “Arounde esturnel et filandre /​Swalewe and starlinge” (Swallow, starling and ‘filandre’) (ll. 809–​810); arounde ‘swallow’ is rendered by ME swalew and esturnel ‘starling’ by ME starling, while filandre is left without a gloss. The compiler might have perceived a similarity between two of the birds of line 809 and left the latter unglossed. It should be highlighted that the Nominale employs the word filandre in another passage (ll. 622–​625) –​without a counterpart in the Tretiz –​, where Bibbesworth juxtaposes the verb form perer(e) ‘to weave’ and the noun perer(e) ‘pear tree’. The same lines bring together another homonymic pair and contrast filandre as ‘starling’ to filaundre as ‘web, cloth’,35 Et de sa cov prist vn cyroyne /​Toke a hande-​worme Graunt merueil fust vn filandre perere /​A sterling warpe Vn teile de filaundre sur vn perere /​A web of gossamer (ll. 620–​625) ‘and she seized an itch-​mite from her tail [the talk is about a mermaid]. /​It was a great marvel that a starling wove /​a cloth36 of gossamer on a pear tree.’

These lines offer proof for a continued penchant for juxtaposing homonyms and confirm the circulation of the pair fila(u)ndre ‘starling’ /​‘web, cloth’.37 This time the threads are woven into a cloth. A weaving loom of pear tree was mentioned in a previous passage the Nominale: “Homme poet teil perer M. may a webbe 34 Cambridge Nom. line 622; Femina (Rothwell 2005: 55). See below for the Westminster glossary; the Add. 8870 Nominale, and the St John’s glossary. 35 For the “homonymic clash” in the Tretiz, see Rothwell (1994). 36 The AND surmises for this occurrence of teile the meaning ‘spider’s web’. But see below. 37 As to several scholars such as Kock (1934: 65 and 73) and Rothwell (2009: 32, note 16 and 79, note 3), the gloss stare is a mistake prompted by the jocular use of filandre in line 719.

74

Patrizia Lendinara

warpe /​Sur paleis fait de perer On a stake of pere-​tre” (One may weave a cloth on a framework of pear-​tree wood) (ll. 335–​338). The ironic turn of these and the former lines is evident as well as the nonsense of a starling weaving a gossamer (line 622). Again, the meaning of gossamer is far from clear. The two further occurrences of fila(u)ndre ‘starling’ are drawn from the Tretiz: in the trilingual Westminster glossary the entry: “philandra: filandre, a starc” (l. stare) occurs in the bird section and follows chalandra (as for other glossary entries, the Latin lemmata seem back-​formations from the French words). In the Add. 8870 Nominale, the lemma is left unglossed, “un fylaundre, –​” (no. 315), but the borrowing from Bibbesworth is proved by its occurrence within the bird section and before the entries un chalaundre and un salamandre. Finally, the St John’s glossary employs the other ME gloss, gossomer, but its entry: “Filaundre . Gossumer .” (no. 36) is found within a sequence of bird names. By the same token, it is on the heels of the occurrence in line 124 of the Tretiz and its gloss, that a series of English glossaries were to include an entry such as “filandr(i)a/​ filaundre: gossomer”. As far as the glossaries compiled after the Tretiz are concerned, an alphabetical glossary such as that in Cambridge, TC, O.5.4 has “Filiandra gosesomere” (ed. WW 583, 16). Some interesting cues are provided by the class-​glossaries where gossamer is listed within atmospheric phenomena or the like. In London, BL, Add. 37075 (f. 315a), “filandria, gossomer” (Ross –​Brooks 1984: 79) occurs within the section bearing the title: ‘Nomina terrarum plantarum [read planetarum] et aliarum rerum’, between “Pruina, hore froste” and “Fons, welle”. In the so-​called Mayer Nominale (now London, BL, Add. 34276: WW 736, 30), “Hoc filandrum a gossomyre” is listed in a batch of atmospheric events. The same happens in the pictorial vocabulary, now Yale University, Beinecke 594: “Hec filandra a gossummer” (WW 802, 17) (the rubric of this section is “Nomina planetarum”). Finally, the Nominale in London, BL, Harley 1002 (ff. 139r–​154r) has “hec ffiliandra anglice gossamer”. The entry is on f. 149r at the end of the section “Nomina terrarum et planetarum” which deals with everything from the planets to weather conditions: the lemma is copied beside entries which deal with the four seasons.38 All these compilations seem to rank gossamer as a seasonal phenomenon because cobwebs are seen floating in the air especially in autumn (more below on the etymology of gossamer). If we take into examination the first dictionaries, the Promptorium parvulorum still combines gossomer with filandrya: “Gossomer, corrupcyon: Filandrya,

38 I would like to thank Annette Horn for her information about the Harley 1002 glossary.

On the Trail of Bibbesworth

75

lanugo, Cath.” (Mayhew 1908 [1973]: 196).39 It is the later dictionary of Palsgrave that records the term in its current meaning: “Gossommer, thynges that flye in sommar lyke copwebbes” (Palsgrave 1530 [1969]: 226). The definition of Cotgrave is similar: “Couvrailles f.: Gossymeare; or the white, and cobweblike exhalations which flye abroad in hot Sunnie weather”.40 The etymology of English gossamer is unknown:41 a composition of goose and summer is defended on the assumption that goose refers to the “downy” appearance of gossamer and that it is applied to the period (St Martin’s summer), when geese were plucked and eaten, cf. German Gänsemonat ‘geese-​month’ (November).42 As for summer, it is thought to refer to a warm spell in autumn, and, in this interpretation, the name of the season was transferred to the spider webs, which are a characteristic phenomenon of the period. Gossamer has also been explained as a loanword from Latin gossympinus ‘the cotton-​tree, the plant that bears cotton’ (Pliny, Nat. Hist. 12,10,21, § 39 and 19,1,2, § 14), Medieval Latin gossypium, gosipium, gossipium. This is the etymology of gossamer accepted by Samuel Johnson in his dictionary. However, words such as French gossampine (Rabelais), denoting a fabric, and English gossampine, with the same meaning, are much later than the Tretiz and all its manuscripts. Gossampine is recorded from the Tudor period and is a loanword from French, born in the context of transatlantic travels and trade.

39 The interpretation refers to the supposition that gossamer was formed from the dew scorched by the morning sun, and hence is termed ‘corruption’. The reference is to the Catholicon of John of Genoa or Johannes Balbus, s.v. Lanugo: “In ‘Lanugo ginis . fe. ge i. lana super poma, vel flos tribuli qui postquam bene siccatus est levi flatu effertur in aerem. et dicitur a lana quia colores lane et est levis ad modum lane […] lanugo est que a vento tollitur”. 40 Filanders (Cotgrave 1611: n.p.) is only registered with the meaning ‘small worms in hawks’: “Filandres f.: the Felanders; small wormes that bread in bruised, surfeited, or foul-​fed, hawkes; also nets to catch wild beast with”. See also the entry: “Filandré m. /​ -​ée f.: Streaked with, or full of, small threads, fibers; felanders”. 41 MED G 3, p. 259: g‫ׅ‬ọ̄s-​somer. [From g‫ׅ‬ọ̄s & sǒmer; cp. MnE dial. go-​summer latter end of summer, summer goose gossamer]; OED s.v.v. gossamer and † go-​summer. 42 ME gossomer is compared to a series of compounds and phrases containing the word for ‘summer’: German Mädchensommer (lit. ‘girls’ summer’), Altweibersommer (‘old women’s summer’), and Sommerfäden. Dutch zomerdraden, Swedish sommartråd, all literally ‘summer thread’. In Germany this time of year is also called fliegender Sommer. I find it difficult to see a connection to gossamer. The obsolete MnE dial. go-​summer and summer-​goose (which is interpreted as a transposition) are also obscure.

76

Patrizia Lendinara

Indeed, ME gossomer might originally denote the flying filaments of rootless twining herbs of a plant such as the cudweed, with its leaves white and woolly, or even the waste of wool, hence its matching with Latin filandr(i)a.43 Except for the occurrences in the Tretiz-​cluster texts, which account also for the occurrences in the fifteenth century glossaries, the first writer to use gossomer is Chaucer in a passage of The Squire’s Tale, where the author vouches its skepticism about atmospheric phenomena presented as marvels, Therfore cesseth hir janglyng and hir wonder. As soore wondren somme on cause of thonder On ebbe, on flood, on gossomer, and on myst, And alle thyng, til that the cause is wyst. (ll. 257–​260) ‘Therefore they stop their chattering and their wonder. /​As intensely some wonder about the cause of thunder, /​of ebb, of flood, of gossamer, of mist, /​and each thing, until the cause is known.’

Chaucer explains that thunder is not caused by demons, but the result of natural actions, which may also explain the nature of gossamer, which can suddenly cover fields and hedges at a certain season of the year. John Lydgate is a writer who is fond of the word gossomer, which he will repeatedly employ in his works.44 In the same vein as Bibbesworth, Lydgate compares gossamer to gold: “Tween gold and gossomer is greet difference” (Halliwell 1840: 46). The matching with gold (a word also linked to gossomer by assonance) occurs in another poem: “Thyng countirfeet hath noon existence;/​ Tween gold and gossomer is greet difference;” (Halliwell 1840: 166). Elsewhere Lydgate condemns someone who would gather gossamer instead of wool, “And gadrithe hym gossomer to pak it for wulle” (Halliwell 1840: 169).45 In the Morte Darthur of Thomas Malory (Krishna 1976: 113), the struck-​ down Gawain compares the fleeting human life, which, in his words, is given to man on deposit, to a gossamer (line 2687), “Greue ȝow noghte,” quod Gawayne, “for Godis luffe of Heuen; For this es bot gosomer and gyffen on erles;

43 For two instances of specialized meaning, see Italian filandra ‘seaweed attached to a ship’s bottom’; filandra, filaria ‘intestinal worms that afflicted hawks’. 44 “Smothe as [vr. as a] gossomer in the hayr”. (Furnivall 1899–​1904. II: 303 l. 11078); “As any gossamer the counterpiece was light” (Lydgate’s application to the Duke of Gloucester for money 4, 1) (Halliwell 1840: 50). The meaning ‘anything light or flimsy’, ‘anything worthless’ is generally dated to ca. 1400, although it was somehow implicit in the Tretiz. 45 For this passage, see Schieberle (2014: 204–​227).

On the Trail of Bibbesworth

77

‘Do not grieve’ said Gawain, ‘for God’s love on Heaven; /​for this is but gossamer and laid on for the time being;’

The word will occur twice in Shakespeare; in Romeo and Juliet ii, 6, 18, ephemerality characterizes both the floating cobwebs and the men who stumble upon them, A lover may bestride the gossamer That idles in the wanton summer air, And yet not fall, so light is vanity.

The word also occurs in King Lear (iv, 6, 49), where Edgar pretends to have seen Gloucester fall from the top of the cliff, saying “Hadst thou been aught but gossamer, feathers, air”. In this case, gossamer is expressive of extreme lightness. These quotations witness a rapid success of the word, which is used in meaning ‘filmy substance consisting of cobwebs’ by Chaucer, who is also responsible for ranking gossamer within atmospheric phenomena. That this was the meaning at the time of the first Bibbesworth glosses is not that certain in my opinion.

6. The Tretiz and the Orthographia Gallica The knowledge of the Tretiz is evident in grammatical texts teaching French through the medium of Latin. The Orthographia gallica, a tract containing rules on the writing of French, comments on its pronunciation and some passages on verbal morphology and gender, quotes several homonyms and near-​homonyms employed by Bibbesworth. Concerning the lists of French words occurring in this work, the editor of the Orthographia writes that: “The lists are plainly compiled with indebtedness to other writers, e. g. Walter of Bibbesworth” (Johnston 1987: 26). The use of the Tretiz is evident in the insistence on homonyms and near-​homonyms such as kyvil ‘pin, peg’ and kevil ‘ankle’ (for which cf. Tretiz line 152).46 When the Orthographia comes to explain the case in which different French words are expressed by a single English word, it resorts to a renowned example of the Tretiz,

46 Johnston (1987: 17). See also the long list in L 50 (p. 15) (“Item diversitas scripture facit differenciam aliquarum diccionum quamvis in voce sint consimiles, verbi gracia […]”) which includes several words occurring in the Tretiz such as rastel ‘rake’ /​rastuer ‘scraper’ (Tretiz ll. 391/​395 G, 306 T), or litter /​littere –​in this case the play is on littere as ‘straw’ and ‘litter’ –​(ll. 397–​402 G, 313–​317 T).

78

Patrizia Lendinara Veez, ci veint devant vous Un chivaler bieau tut rous Qui une destrere sor se est munté. (reed) Esku de goules ad porté, (reed) Un launce rouge en l’uyn mein, De vin vermaille l’autre plein, Qi ne manjuwe point de peschoun S[i]‌de le haranc sor noun. (reed) (ll. 310–​317) ‘Look, here comes up to you /​a handsome red-​haired knight, /​mounted on a red [= chestnut] horse. /​He carries a red shield, /​a red lance in one hand, /​the other full of red wine; /​he eats no fish but red [= smoked] herring.’

The presentation is much reduced: “Item habentur diversa verba Gallica pro isto verbo anglico ‘reed’, videlicet rous chivaler, chival; et harang soor; escue de goules; une rose vermaile” (L 87), but it vouches unmistakably for the popularity of the Tretiz (Johnston 1987: 17).47 Also the dialogues of the fifteenth-​century Liber Donati, which contains lists of birds, animals, parts of the body, clothing, etc., are very reminiscent of the corresponding lists in Bibbesworth. Again, the additional lines of the manuscripts of the Tretiz increase the number of correspondences between the two works.48 The Liber is in large part devoted to giving French equivalents of Latin verbs and adverbs, followed by the names of the days of the week and numbers, and, finally, by seven pages of the so-​called dialogues. A second category of teaching manuals, which were quite popular in the later fourteenth and early fifteenth centuries, is that of the Manières de langage. These texts were produced for English travellers going to France, but the destination of their drills might not be that exclusive, and the manuals of conversation in French might have been destined for a larger audience. The Manières consist of dialogues in French between the traveller and all the different people he is supposed to meet, with long strings of terms belonging to specific areas of activity. A part of the imaginary exchanges covers the same ground as the Tretiz and incorporate batches of words employed by Bibbesworth.

47 For a verb, see L 88 (p. 18) “item habentur diversa verba Gallica pro isto verbo Anglico ‘breke’ […]” and Tretiz ll. 1059–​1062 G; 782–​785 T (accompanied by relevant glosses in T) for the four A-​N verbs (fruisser, debriser, rumper and enfeindre) that can be translated by ME brēken. 48 For example, cf. Merrilees –​Sitarz-​Fitzpatrick (1993: 24–​25) for the overlapping of the section on the parts of the body.

On the Trail of Bibbesworth

79

In the Manière de langage (1415) the traveller asks the innkeeper about dinner. The mistress replies with a long list of volatiles and wild birds to cook for dinner, followed by a list of tableware, tablecloths and house implements. Many of the terms were at hand in the Tretiz and were drawn from its section on arranging a feast (ll. 1105–​1140 G; ll. 812–​838 T). In the following exchanges of the Manière (1415) there occur a list of bedding or bed-​related implements, a second list of food, and a catalogue of animals and other items to be sold at the market by the mistress’ husband, including different kinds of wool fabrics and clothes. The son of the mistress, called to the scene and questioned about his level of instruction, rattles off a list of the parts of the body and another of garments (Kristol 1995: 76–​78). These last, in particular, are very close to the Tretiz. The description of the parts of the body is an abbreviated version of the corresponding section of the Tretiz (ll. 29–​182). There follows a longer list of sea food and fish to eat the following day before the departure. The long list of fish names in the Manière (1415) resembles the one given in the B, O, 4 manuscripts of Bibbesworth (Kristol 1995: 43–​44). On the contrary, the first of the Extraits de Nominalia printed by Owen as an appendix to her edition has no relationship with the Tretiz. The piece on p. 145 is drawn from Oxford, Magdalen College 188, f. 5v (“apparatus pro corpore”) and belongs to a household account.49 Tracts on household accounts might be written in either Latin or French and are de facto intended to list house and farm implements, clothing and all sort of goods in detail. In more than one instance these inventories were turned into a glossary by means of the addition of bilingual counterparts. In London, BL, Harley 4971 (s. xiv3/​4), a household account in French (ff. 6v–​9r) includes inventories of goods and products; in the last folios (ff. 8r–​9r) the list becomes bilingual (French-​English) and includes headings in French (e.g. Ore des vifs oiselx). The same is true of Oxford, Magdalen College, 188 (s. xv2/​4), where, on ff. 5r–​7v, a series of trilingual batches (Latin-​French-​English) are arranged under Latin headings. Although the distinction, as in this case, is a fine one, these texts should be kept distinct from learning tools given that they perform a practical function. The similarity in the order of the household goods listed in the accounts may only be fortuitous.50 49 The other excerpts printed by Owen (1929: 146–​149), which are often referred to incorrectly, are drawn from the Manière de langage (1399) (in Oxford, All Souls College 182), the Cambridge Nominale, Femina and its final tripartite guide (this fourth section of Femina is based on the first one). 50 For an inspiring scrutiny of this genre of texts, see Smith (2003). A parallel with the Tretiz is surmised at Smith (2003: 229, note 45).

80

Patrizia Lendinara

7. Further Possible Influence The Tretiz, owing to its large popularity and circulation, might also have exerted its influence on literary compositions. The Anglo-​Norman Geste de Blancheflour et de Florence features a number of lists of words belonging to the same lexical field and also betrays a strong penchant for enumeration.51 The couplets nos. 3–​6 of the Geste list musical instruments; nos. 7–​9 precious stones; nos. 10–​12 trees; and nos. 13–​15 birds. The parallels with the Tretiz are frequent in the last two groups of couplets, but the date suggested for the Geste –​the late thirteenth century –​is quite close to that of the Tretiz and, possibly, the Geste rather witnesses to the contemporary demand for French teaching also with the help of a playful strings of words. Dominica Legge (1963: 334–​335) suggested that one of the aims of the Geste may have been to teach French vocabulary. The circulation of compilations with a similar aim is witnessed by the other lexical fields exploited in the Geste, such as that of precious stones.52 Similarly, the lay-​out of these verses, which tend to have three terms (and sometimes four) a line, reminds us of the Tretiz (where the cumulative effect is enhanced by the use of tail-​rhyme stanzas instead of couplets); compare, for example, “Ceriser, pyne e coygner” (Geste line 62) and “Pomer, perere e cereise” (Tretiz line 677). The trend to list tens of birds, trees or food, with either an erudite, didactic or humorous intent, was destined to endure in English literature. In the Knight’s Tale, in describing Arcite’s funeral, Chaucer implements the catalogue of Boccaccio, adding ook, birch, aspe, holm, popler, wylugh, plane, box, chasteyn, lynde, laurer, mapul, thorn bech, ew, and whippeltree (ll. 2921–​2923).53 Many of these tree names were already there in the glosses to the Tretiz.

51 See Meyer (1908b: 222 and 224) for comments on the lexicon of the poem. The Old French romance Floire et Blancheflor (composed ca. 1160–​1170) (Leclanche 1980) has short lists of fruits (ll. 1687–​1689), drink and food, including game (ll. 1673–​1683), birds (ll. 2000–​2004), precious stones (ll. 2015–​2019), and game birds (ll. 3185–​3194). None bears comparison with those of La Geste de Blancheflour et de Florence. The flowers and birds of the garden of delights receive a joint description and are not presented as a list (ll. 2021–​2044). The two poems are also quite different in tone, structure and ending. Melior et Ydoine (Meyer 1908b) features no list whatsoever. As to the ME version, Floris and Blancheflour, this has only a small cluster of precious stone names in the description of the orchard of the emir (ll. 662–​663). 52 For precious stones, see, e.g. ll. 112–​124 of the poem How a Lover Praiseth his Lady (Hammond 1923–​1924). The poem has lists of birds on ll. 101–​110 and 148–​149. 53 See Schildgen (2013: 96). Chaucer added the names of native species of trees.

On the Trail of Bibbesworth

81

8. Conclusions While two surviving texts (the Cambridge Nominale and Femina) offer abbreviated versions of the Tretiz, Bibbesworth’s poem is drawn on by a number of glossaries and by works employed for formal teaching of French language. All these compositions are mainly interested in the rich lexicon of the Tretiz and, as far as bilingual (and trilingual) glossaries are concerned, willing to take advantage of the ready-​made combination of French words and ME glosses offered by the several manuscripts of the Tretiz. The many subtleties a multi-​layered text such as the Tretiz are inevitably lost in these borrowings, first among them the wordplay which characterizes many lines of the original work. These fourteenth-​ and fifteenth century texts were designed for a different milieu and different and likely less developed learners of French language. As such this reusing of a mid-​ thirteenth century composition yields remarkable data on the evolution of the approach to language teaching in medieval England.

References Primary sources (including earlier dictionaries) Cecchini, Enzo et al. (eds.). 2004. Uguccione da Pisa: Derivationes. 2 vols. Firenze: SISMEL Edizioni del Galluzzo. Cotgrave, Randle (ed.). 1611. A Dictionarie of the French and English Tongues. London: Adam Islip. http://​www.pbm.com/​~lindahl/​cotgrave/​ [last accessed 20 December 2020]. Furnivall, Frederick J. (ed.). 1899–​1904. The Pilgrimage of the Life of Man, Translated by John Lydgate. (EETS ES 77, 83, 92). London: Kegan Paul, Trench, Trübner. Halliwell, James O. (ed.). 1840. A Selection from the Minor Poems of Dan John Lydgate. London: Percy Society by C. Richards. Johnston, R. C. 1987. Orthographia Gallica. (ANTS Plain Texts Series 5). London: ANTS. Krishna, Valerie (ed.). 1976. The Alliterative Morte Arthur. New York: Burt Franklin. Kristol, Andres M. (ed.). 1995. Manières de Langage (1396, 1399, 1415). London: ANTS. Legge, M. Dominica (ed.). 1941. Anglo-​Norman Letters and Petitions from All Souls Ms. 182. (ANTS 3). Oxford: Blackwell.

82

Patrizia Lendinara

Leclanche, Jean-​Luc (ed.). 1980. Le conte de Floire et Blanchefleur. Roman pré-​ courtois du milieu du XIIe siècle. Paris: Champion. Mayhew, A. L. (ed.). 1908. The Promptorium Parvulorum: The First English-​ Latin Dictionary, EETS ES 102. London: Oxford University Press. Repr. 1973. Millwood /​New York: Kraus. Merrilees, Brian –​Beata Sitarz-​Fitzpatrick (eds.). 1993. Liber Donati. A Fifteenth Century Manual of French. London: ANTS. Mowat, J. L. G. (ed.). 1882. Sinonoma Bartholomei. Oxford: Clarendon Press. Owen, Annie (ed.). 1929. Le traité de Walter de Bibbesworth sur la langue française. Paris: Les Presses Universitaires. Palsgrave, John (ed.). 1530. L’ éclaircissement de la langue française [Lesclarcissement de la langue francoyse]. Finished by Johan Haukyns. London: Richard Pynson [facsimile edition 1969 Menston: Scolar Press]. Rindal, Magnus (ed.). 1981. Barlaams ok Josaphats saga. Oslo: Norsk historisk kjeldeskrift-​institutt. Ross, Thomas W. –​Edward Brooks Jr (eds.) 1984. English Glosses from British Library Additional Manuscript 37075. Norman, OK: Pilgrim Books. Rothwell, William (ed.). 2005. Femina (Trinity College, Cambridge MS B.14.40). The Anglo-​Norman On-​Line Hub, Texts and Publications. Aberystwyth /​ Swansea. The Anglo-​ Norman Hub. http ://​www.anglo-​norman.net/​texts/​ femina.pdf [last accessed 20 December 2020]. Rothwell, William (ed.). 2009. Walter de Bibbesworth, Le Tretiz, from MS. G (Cambridge University Library Gg.1.1) and MS. T (Trinity College, Cambridge O.2.21) together with Two Anglo-​French Poems in Praise of Women (British Library, MS. Additional 46919). The Anglo-​Norman On-​Line Hub, Texts and Publications. Aberystwyth /​Swansea: The Anglo-​Norman Hub. http ://​www. anglo-​norman.net/​texts/​ [last accessed 20 December 2020] Skeat, W. W. (ed.). 1903–​1906. “Nominale sive Verbale”. Transactions of the Philological Society 1903–​1906, pp. 1*–​50*. WW= Wright, Thomas –​Richard P. Wülcker (eds.). 18842. Anglo-​Saxon and Old English Vocabularies. London: Trübner.

Secondary literature Acker, Paul. 1993. “An Anglo-​Norman-​Middle English Glossary of Tree and Bird Names”. Medium Ævum 62: 285–​288. Aldís Sigurðardóttir et al. (eds.). 1995–​. Ordbog over det norrøne prosasprog /​A Dictionary of Old Norse Prose. Copenhagen: Den Arnamagnæanske Kommission. http://​onp.ku.dk/​onp/​ [accessed 20 December 2020]

On the Trail of Bibbesworth

83

AND = Rothwell, William et al. (eds). 1977–​19922. Anglo-​Norman Dictionary. London: MHRA. http://​www.anglo-​norman.net [accessed 20 December 2020]. Baker, J. H. 1989. “A French Vocabulary and Conversation-​Guide in a Fifteenth-​ Century Legal Notebook”. Medium Ævum 58: 79–​102. Baugh, Albert C. 1959. “The Date of Walter of Bibbesworth’s Traité”. In: Oppel, Horst (ed.): pp. 21–​33. Biggam, Carol P. (ed.). 2003. From Earth to Art: The Many Aspects of the Plant-​ World in Anglo-​Saxon England. Proceedings of the First ASPNS Symposium, University of Glasgow, 5–​7 Apr. 2000. Amsterdam: Rodopi. Björkman, Erik. 1900–​1902. Scandinavian Loanwords in Middle English. 2 vols. Halle a. S.: Karras. Classen, Albrecht (ed.). 2005. Childhood in the Middle Ages and the Renaissance. Berlin: de Gruyter. Cheney, C. 1972. “Law and Letters in Fourteenth Century Durham: A Study of Corpus Christi College, Cambridge, Ms. 450”. Bulletin of the John Rylands Library 55/​1: 60–​85. Cleasby, Richard –​Gudbrand Vigfusson (eds.). 19572. An Icelandic-​English Dictionary. With a Supplement by William A. Craigie. Oxford: Clarendon Press. http://​www.ling.upenn.edu/​~kurisuto/​germanic/​oi_​cleasbyvigfusson_​ about.html [accessed 20 December 2020]. Dearnley, Elizabeth. 2016. Translators and Their Prologues in Medieval England. Cambridge: Boydell & Brewer. Evans, Dafydd. 1993. “The Taxonomy of Bird-​Naming in Anglo-​Norman and in Channel Island Patois”. In: Short, Ian (ed.): 105–​134. FEW= von Wartburg, Walther (ed.). 1928–​. Französisches etymologisches Wörterbuch. Bonn: Klopp. Hammond, Eleanor P. 1923–​1924. “How a Lover Praiseth his Lady”. Modern Philology 21: 379–​395. Hinton, Thomas. 2017. “Anglo-​French in the Thirteenth Century: A Reappraisal of Walter de Bibbesworth’s Tretiz”. Modern Language Review 112: 848–​874. Hough, Carole, ‘Place-​ Name Evidence for Anglo-​ Saxon Plant-​ Names”. In: Biggam, Carol P. (ed.): 41–​78. Hunt, Tony. 1979. “Les gloses en langue vulgaire dans les Mss de l’Unum omnium de Jean de Garlande”. Revue de Linguistique Romane 43: 162–​178. Hunt, Tony. 1981. “The Trilingual Vocabulary in MS Westminster Abbey 34/​11”. Notes and Queries 226: 14–​15. Hunt, Tony. 1991. Teaching and Learning Latin in Thirteenth-​Century England. 3 vols. Cambridge: Brewer.

84

Patrizia Lendinara

Hunt, Tony. 1993. “Anglo-​Norman Medical Receipts”. In: Short, Ian (ed.): 179–​233. Ingham, Richard. 2014. “The Maintenance of French in Later Medieval England” Neuphilologische Mitteilungen 115: 425–​448. Jambeck, Karen K. 2005. “The Tretiz of Walter of Bibbesworth: Cultivating the Vernacular”. In: Classen, Albrecht (ed.): 159–​184. James, Montague Rodes. 1913. A Descriptive Catalogue of the Manuscripts in the Library of St. John’s College. Cambridge: University Press. Kitson, Peter R. 1997. “Old English Bird-​Names. I”. English Studies 78: 481–​505. Knox, Philip. 2013. “The English Glosses in Walter of Bibbesworth’s Tretiz”. Notes and Queries 60: 349–​359. Koch, John. 1934. “Der anglonormannische Traktat des Walter von Bibbesworth in seiner Bedeutung für die Anglistik”. Anglia 58: 30–​77. Kries, Susanne. 2003. Skandinavisch-​ schottische Sprachbeziehungen im Mittelalter. I. Der altnordische Lehneinfluss. Odense: University Press of Southern Denmark. Laing, Margaret. 2003. Catalogue of Sources for a Linguistic Atlas of Early Medieval English. Cambridge: Boydell & Brewer. Legge, M. Dominica. 1963. Anglo-​ Norman Literature and its Background. Oxford: Clarendon Press. Lendinara, Patrizia. 2015. “The Glossary in ms. Cambridge, St John’s College E.17 and Middle English Lexicography”. Filologia Germanica –​Germanic Philology 7: 89–​140. Lockwood, W. B. 1984. The Oxford Book of British Bird Names. Oxford /​ New York: Oxford University Press. MED = Kurath, Hans –​Sherman M. Kuhn et al. 1952–​2001. Middle English Dictionary. Ann Arbor, Michigan: University of Michigan Press. Online edition in McSparran, Frances et al. (eds.). 2000–​2018. Middle English Compendium. Ann Arbor, MI: University of Michigan Library. https://​quod.lib.umich.edu/​ m/​middle-​english-​dictionary/​dictionary [accessed 20 December 2020]. Meyer, Paul. 1908a. “Notice du ms. 25970 de la Bibliothèque Phillipps (Cheltenham)”. Romania 37: 209–​235. Meyer, Paul. 1908b. “Melior et Ydoine”. Romania 37: 236–​244. OED= Oxford English Dictionary Online. Oxford University Press. http://​www. oed.com [accessed 20 December 2020]. Oppel, Horst (ed.). 1959. Festschrift für Walther Fischer. Heidelberg: Winter. Pilcher, Jeffrey M. 2012. The Oxford Handbook of Food History. New York /​ Oxford: Oxford University Press.

On the Trail of Bibbesworth

85

Rothwell, William. 1982. “A Mis-​judged Author and a Mis-​used Text: Walter de Bibbesworth and his ‘Tretiz’ ”. Modern Language Review 77: 282–​293. Rothwell, William. 1994. “Of Kings and Queens, or Nets and Frogs: Anglo-​ French Homonymics”. French Studies 48: 257–​273. Schieberle, Misty. 2014. “Proverbial Fools and Rival Wisdom: Lydgate’s Order of Fools and Marcolf ”. The Chaucer Review 49: 204–​227. Schildgen, Brenda Deen. 2013. “Reception, Elegy, and Eco-​Awareness: Trees in Statius, Boccaccio, and Chaucer”. Comparative Literature 65: 85–​100. Short, Ian (ed.). 1993. Anglo-​Norman Anniversary Essays. (ANTS. Occasional Publications Series 2). London: ANTS. Smith, D. Vance. 2003. Arts of Possession. The Middle English Household Imaginary. (Medieval Cultures 33). Minneapolis: University of Minnesota Press. Thurot, Ch. 1986. “Notices et extraits de divers manuscrits latins pour servir a l’histoire des doctrines grammaticales au moyen âge”. Notices et extraits des manuscrits de la Bibliothèque Impériale et autres bibliothèques 22,2 (Paris: Imprimerie Impériale, 1868): 1–​540. Thomson, Rodney M. 2013. Catalogue of Medieval Manuscripts of Latin Commentaries on Aristotle in British Libraries, II. Cambridge. Turnhout: Brepols. TLF = 1971–​1994. Trésor de la langue française. 16 vols. Nancy /​Paris: édition du CNRS/​Gallimard). http://​atilf.atilf.fr/​ [accessed 20 December 2020]. Wright, Joseph (ed.). 1898–​ 1905. The English Dialect Dictionary. 6 vols. Oxford: Henry Frowde.

Alessandra Petrina

The Construction of the European Intellectual: Petrarch, Humanism, and Middle English Literature Abstract: Petrarch’s reflections on fame and the legacy of classical tradition prompted the inscription of poetry within the wider structure of human history. He strove to comprehend a development of culture that clamoured to be understood in its own terms, beyond the overarching reference to the divine plan. This sometimes painful search brought him to be hailed, in centuries to come, as a proto-​humanist writer. As late-​medieval English literature struggled to find its identity, in linguistic and cultural terms, the legacy of Petrarch proved essential, durable, and complex. The Petrarchan texts drawn upon and the reactions they generated changed, sometimes radically, providing a singular instance of translatio studii: translations, rewritings, and parodies from Petrarch chart the passage of English writing from the late Middle Ages to the Renaissance. Keywords: Petrarch, Geoffrey Chaucer, John Lydgate, Winchester Anthology

“Rara lectio est quae periculo vacet” (Dotti 1991:90) –​reading rarely avoids danger (Kahn 1985: 154), as Petrarch writes in one of his epistles. In the attempt to read Petrarch’s role in the English Middle Ages the danger is of falling captive to that most complex of writers, Petrarch himself. My investigation of the role played by Petrarch in the English Middle Ages focuses on three instances: Geoffrey Chaucer, John Lydgate, and an anonymous manuscript compilation known as the Winchester Anthology. The contacts between Petrarch and England, while the poet was alive, were minimal, and on the poet’s part appear to have focussed on geographical dislocation. In one of his letters, addressed to Tommaso Caloiro da Messina, Petrarch harks back to Pomponius Mela’s Cosmographia, repeating its claim that the island is shaped like Sicily: “Hoccine est quod in Cosmographia Pomponii legeram, Sicilie maxime similem esse Britanniam?” (Dotti 1991: 50).1 More interesting, in another letter to the same Tommaso, is Petrarch’s description of an encounter 1 ‘Is it for this that Pomponius maintains in his Cosmographia, that Britannia is especially similar to Sicily?’. The allusion is to Pomponius Mela’s De Chorographia III.50. Translations are mine unless otherwise indicated.

88

Alessandra Petrina

with Richard of Bury. The letter refers to the trip to northern Europe Petrarch had undertaken in the spring of 1333 (Collette 2018). The supposed encounter with Richard, who promises to give Petrarch information about the location of the island of Thule, and then never answers his questions, is a pretext for Petrarch to evoke an unexplored place, belonging more to imagination than to geography: “Perambulanti veterum confinia, accessu quidem aspera sed amena cum perveneris, creber scrupulus ingenii pede calcandus est” (Dotti 1994: 2).2 But the letter presents a number of problems: it was probably written much later than its fictional date of 1333, and Tommaso Caloiro may never have read it (Cachey 2003: 77–​84). On the other hand, in none of his writings does Richard of Bury mention meeting Petrarch. We can read this episode simply as a first, aborted attempt at contact between English and Italian culture; or as a complex metaphor Petrarch constructs to use geography as a paradigm of the pilgrimages of the mind. It is something that will be familiar to us in the following centuries –​as shown, for instance, by Thomas More when creating his island of Utopia, or by John Donne ecstatically exclaiming, of the body of his lover, “O my America, my new found land”. But at this early stage Petrarch was anticipating his contemporaries in exploring the boundaries of knowledge, through books and maps; reading and prima facie experience are uncannily fused in his letter, which focuses on the ultimately impossible search for the ultima Thule: “Quero, sed, ut verum fatear, nec certo indice nec ad rem ducentibus coniecturis, aut ipsam aut ullam inveniendi spem invenio. Et hec tibi quidem, ex ipsis britannici occeani litoribus, propinquior –​ut fama est –​ipsi quam vestigamus insule, scribo; profecto unde, vel antiquo literarum studio vel nova ac solicita locorum indagine, certius aliquid scribere posse debueram. Ultimam quippe terrarum esse, non ambigitur: hoc Virgilium canit, hoc Seneca, hoc secutus utrunque Boetius, hoc omnis denique scriptorum cohors” (Dotti 1994: 2).3 The poet’s quest crosses the boundaries between what is read and what is experienced; the true explorers are no longer the cosmographers, like Pomponio Mela, but the 2 ‘For him who travels across the boundaries of the ancients, as he arrives to places that are hard to reach yet pleasant, the frequent effort of the mind must be employed’. 3 ‘I seek it but, to tell the truth, without a clear clue or something that might lead me to a conjecture; and I have no hope of finding it, or any other island. And I am writing this to you from the very shores of the British ocean –​quite close, or so they say, to the very island we are investigating; starting from here, either by means of ancient studies or of an immediate search in the place, I should have been able to write something more certain. There is no doubt that it is the farthest land: so sings Virgil, so Seneca, so –​following both –​Boethius, so the cohort of all other writers’.

The Construction of the European Intellectual

89

poets-​philosophers, from Virgil to Seneca, to Boethius, who follows them both. What Petrarch is proposing is a journey across the landscape of writing. The evocation of the great writers of the past is a topos that had already found a great model in Dante’s Inferno IV, as the poet found himself in Limbo, in the company of the great poets of the classical past: the evocation of the classical canon served also to enclose the writer as a latter addition to the canon. Dante is quite explicit about this, suggesting that in the gathering of the five poets he became the sixth; Petrarch is more wistful in the passage above, suggesting a search “nec certo indice”, without a clear clue. Another exploration appears in the closing scene of Africa, the poem Petrarch considered his masterpiece and on which he worked all his life. A dream vision involves three poets from three different ages: the Greek Homer, the Latin Ennius and himself. Strongly influenced by the Somnium Scipionis, the dream vision appearing in book VI of Cicero’s De re publica which played such an important role in medieval dream visions, this scene is at the same time a summa of the poem and a development of a classical theme in a new, wholly original direction. As Scipio’s interlocutor and poetic voice, Ennius is an overarching presence in the poem; in Africa II.444 Petrarch had already called himself Ennius alter, a second Ennius. In this final book IX Ennius himself dreams under a tree, and sees in his dream Homer, the Greek poet, approaching. This particular dream explores a well-​known literary locus in classical Rome and in the Middle Ages, as it formed the prohemium of Ennius’ own Annales, envisaging a metempsychosis of Homer’s soul into Ennius, justifying and giving authority to Ennius’ whole literary enterprise. Petrarch takes up the suggestion, developing the image still further: the first part of Book IX sets the re-​telling of the dream in a dialogue between Ennius and Scipio Africanus, while the second part, starting from line 216, presents an interesting development: “Hic ego –​nam longe clausa sub valle sedentem /​Aspexi iuvenem” (Lenoir 2002: 430).4 The newcomer is a young man sitting among laurel trees, resembling the depiction of Virgil in Simone Martini’s miniature in the Ambrosianus manuscript (Mann 1984: 106), meditating upon some great project, and at the same time seeking to crown himself with a green branch. Ennius’ vision encompasses not only the young man but also his surroundings: Hunc tibi Tusca dabit latis Florentia muris Romulea radice oriens, urbs inclita quondam,

4 ‘Here I saw a young man sitting for a long time in an enclosed valley’, IX.216–​17. The edition used throughout is Lenoir (2002).

90

Alessandra Petrina Nunc nichil. Utve quaeas ortus confinia nosse, Divitis egregius muros interluet urbis Arnus in Ausonie descendens menia Pise. (IX.224–​28)5

This young man will be called upon to revive the long-​dormant song of the Muses and will re-​tell in one poem all the events Ennius witnessed, thus giving new life to the ancient story. The poem is, of course, Africa, and the young man is “Francisco cui nomen erit” (IX.232), Petrarch himself. More follows: the same young man will make these events, the poem, and himself, eternal by receiving the poetic coronation on the Roman Capitol. More explicit than the passage about Ultima Thule, the scene is typical of Petrarch’s construction of the self, evident also in other texts, from Secretum to Posteritati, but also in extreme and splendid gestures such as the poetic coronation. The latter has also been made the object of some gentle fun on the part of scholars, who have read this as an instance of Petrarch’s obsession with self-​recognition: after some insistent lobbying on his part, the poet was indeed crowned with laurel in the Roman Capitol in 1341 (that is, probably before writing this section of the Africa), but by the time of his coronation his poetic output in Latin was far from impressive, and the epic poem itself, of course, was as yet incomplete. The coronation ceremony may be read as an anticipation of things to come rather than the celebration of an achievement (Wilkins 1943: 168–​70):6 as has been observed, “the ceremony and the oration do not so much recognize Petrarch’s fame as create it. He becomes, in an innovative way, an international celebrity, famous for being famous” (Boswell –​Braden 2012: 2). In all these instances we see the poet striving to define himself, to find his collocation in space (as shown by the identification of Rome as the spring of all poetry) and time (by inscribing his name within a canon that harks back to classical times). But the concluding scene of Africa has also some interesting features that highlight Petrarch’s analysis of the construction of a poetic voice. The first is the fact that, unlike what happens in Ennius’ original, this episode of Africa does not feature metempsychosis. Instead, in detailing Ennius’ evocation of Homer, Petrarch suggests that what the character is experiencing is an inner

5 ‘He will come to you from the Tuscan Florence, a city of wide walls that springs from Romulus’ roots; the city will be one day famous; now it is nothing. So as you may know, the site of his origin, the noble Arno, running along the ramparts of the Ausonian Pisa, shall touch the walls of this rich city’. 6 Wilkins also offers a list of the Latin poems Petrarch had probably published before he was awarded the poetic laurel (1943: 169).

The Construction of the European Intellectual

91

vision: “Presentemque animo ficta sub ymagine feci”.7 Not only does this choice reflect a preoccupation with avoiding the pre-​Christian overtones associated with metempsychosis; it also furthers the exploration of the self that Petrarch had undertaken throughout his writings, most notably in Secretum. Once the fictional image develops further, the dreamer –​Ennius –​offers us a vision not only of the past but also of the future, as underlined by Petrarch’s use of verb tenses. Anagnorisis does not serve the needs of a narrative plot but becomes poetic recognition: by identifying the Franciscus of Ennius’ dream with the Italian poet who is also the material author of the dream, the reader is induced to reflect on the creative ability of poetry. For students of Chaucer there is an immediate link: like the Africa, Chaucer’s dream poems were of course also influenced by the Somnium Scipionis, and the theme of dreaming is central to the meditation on poetic inspiration.8 But another other important link is the Canticus Troili, Chaucer’s translation of Petrarch’s Sonnet 132 from the Canzoniere in Troilus and Criseyde. This is one of the great puzzles for Chaucer scholars, since it offers a completely isolated instance of Petrarch’s Italian poetry entering English literature more than a century before the systematic absorption, through translation, adaptation and imitation, of both Trionfi and Canzoniere. What Chaucer takes from Petrarch is not the facile love convention that will come to be associated with early modern Petrarchism, but a poetic of desire, expressed in Petrarch through the mediation of Augustine, in Chaucer through the mediation of Boethius.9 Petrarch’s sonnet has an importance for Chaucer that goes beyond its belonging to a wider and more articulate context: the sonnet is not only translated in isolation, but the Troilus narrator very explicitly marks a suspension of the narration to insert it (Ginsberg 2011: 121, and also Rossiter 2010: 109–​31). If we recall the quotation from Africa, “Presentemque animo ficta sub ymagine feci”, we find an interesting echo in the Chaucerian construction introducing the translation of Petrarch’s

7 ‘I made him present to my soul as a fictional image’, IX.151. 8 The analogy between Chaucer’s dream poems and Petrarch’s Africa (and possibly Trionfi) has been noted, among others in Andreoni Fontecedro (2011: 345–​48). 9 Chaucer in fact might have been studying Petrarch’s sonnet and Boethius roughly in the same years. See Kaylor Jr. (1993: 221–​22). Chaucer’s “poetic of desire” has been described by John Freccero: “Each of the successive desires of life are in fact desires for selfhood, expressed metonymically in an ascending hierarchy of abstraction: nourishment for the child, sex for the adolescent, fame for the adult. In an Augustinian world, there is no escape from desire short of God: ‘Our heart [he says] is unquiet until it rests in Thee’ [Confessions, I, 1]” (Freccero 1975: 35).

92

Alessandra Petrina

sonnet. Troilus (in a marked deviation from Boccaccio’s original) prepares his mind (as underlined by the use of gan) projecting upon it a vision he is preserving in his heart: Thus gan he make a mirour of his mynde In which he saugh al holly hire figure, And that he wel koude in his herte fynde. (I. 365–​67)10

The image of the mirror, as has been shown, derives from Boethius’ De consolatione philosophiae, V, m. 3, 18–​20: thus, Chaucer inscribes a Boethian motif on the Petrarchan text.11 The lines that follow go back to Boccaccio’s original in declaring Troilus’ readiness to make every effort for his love –​but at this point such efforts can no longer be read in the conventional frame of the lover ready to sacrifice himself or to risk all dangers; the effort is clearly the power of imagination. If Boethius gives Chaucer’s poem its philosophical framework, the influence of Petrarch turns the central metaphor of the poem, the generating power of love (celebrated in the extraordinary prologue to Book 3), into the articulation of a conscious act of literature, as can be seen in the invocation to Venus: Now, lady bryght, for thi benignite, At reverence of hem that serven the, Whos clerc I am, so techeth me devyse Som joye of that is felt in thi servyse. (III. 39–​42)

Such creative force is never underestimated in Chaucer –​it comes back, with a comic twist, in The Miller’s Tale, when the narrator reminds us that “men may dyen of ymaginacioun” (Benson 1988: 73, l. 3612), and it is the object of a deliberate equivocation between the craft of poetry and the craft of love in the opening stanza of The Parliament of Fowls: The lyf so short, the craft so long to lerne, Th’assay so hard, so sharp the conquerynge, The dredful joye alwey that slit so yerne: Al this mene I by Love, that my felynge Astonyeth with his wonderful werkynge So sore, iwis, that whan I on hym thynke Nat wot I wel wher that I flete or synke. (ll. 1–​7)

1 0 All quotations from Chaucer are taken from Benson (1988). 11 For a discussion of this point see D’Agata D’Ottavi (2011: 440). For all quotations from Boethius the edition used is Stewart –​Rand –​S.J. Tester (1973).

The Construction of the European Intellectual

93

Both love and poetry require the concentrated effort of the mind, a labour that is both dedicated and difficult: Chaucer is fascinated by the idea of painstaking craftmanship that is behind it. In Troilus and Criseyde, the allusion to Lollius (explicitly mentioned in the stanza preceding the Canticus) appears to underline the fact that Boccaccio’s original is only partly responsible for the new creation: Troilus and Criseyde proposes itself as a newly-​born creature with a fictional father, a Lollius, who is neither Petrarch nor Boccaccio. The Canticus realizes Alanus ab Insulis’ dictum: “Poësis mentali intellectui materialis vocis mihi depinxit imaginem et quasi archetypa verba idealiter praeconcepta vocaliter produxit in actum” (de Bruyne 1946: 298).12 Troilus’ self-​discovery is effected not only through love, but through creation –​the consequence of love. The connection between Love and the process of knowledge is also underlined in the translation of Petrarch’s opening line: Petrarch’s “S’amor non è, che dunque è quel ch’io sento?” is rendered as “If no love is, O God, what fele I so?” (I. 400). The elimination of Petrarch’s dunque cancels the inference that Love has caused the lover’s present state: as has been observed, “instead of positing Love as the source of his particular state, he begins with what he senses, the starting point, as Aristotle had insisted, of all knowing. As a result, Love is demoted from a priori cause to negative inference: what else could it be that has made him to feel as he does?” (Ginsberg 2011: 123). A side-​by-​side reading of the two poems shows Chaucer not deviating from the Petrarchan original, rather expanding it, exploring it, as if to take it as a discipline, an intellectual or spiritual exercise: the initial exploration of love is quickly turned into an exploration of the self –​the very number of pronouns in the first-​person singular is doubled in Chaucer’s translation. The final dream in Africa also highlights the function of the poetic effort within a community, and this is another important theme that will be central in the reflection of late medieval English poets. Petrarch uses the dream to construct a poetic self that is not simply looking back at a literary past but projecting himself and his works in the future: as we see more clearly in Posteritati, the letter he writes to posterity, Petrarch deliberately creates a literary agenda that involves poets before and after him. In Posteritati, as in the dream closing Africa, prophecy is exploited to construct the role of poetry in history, and to offer indications on the active participation of the intellectual in its progress. Poetry, in these as in other texts, is not simply an antidote to death, but rather projects further and

12 ‘Poetry painted for me an image for the intellect thanks to a material voice, and, almost as archetypes, proposed ideally pre-​conceived words vocally, in deed’.

94

Alessandra Petrina

more terrible possibilities of death: in both Triumphi and Africa we find, after the first, physical death, the possibility of a second death connected to the vanishing of fame (a concept inspired by Boethius’ De consolatione philosophiae, II m. 7, 26): Tutto vince e ritoglie il Tempo avaro; chiamasi Fama, et è morir secondo; né più che contra ’l primo è alcun riparo. (Ariani 1988: 378, Triumphus Temporis ll.142–​44)13

Africa envisages a third death connected to the destruction of the books which only guarantee memory (Andreoni Fontecedro 2011: 341): libris autem morientibus ipse Occumbes etiam; sic mors tibi tertia restat. (II.465–​66)14

It would, however, be a mistake to think that this contemplation of the fragility of our written inheritance simply filled Petrarch with the terrified sense of ubi sunt that is such a strong part of the medieval meditation on the passing of time, and that we find in such early texts as the Old English elegies. In this sense we identify proto-​humanism in Petrarch: an active habit of the mind that reflects on the new role of poetry, and Petrarch’s most enduring legacy to the English Middle Ages. Within this frame we may view the poetic coronation as a revival of a classical tradition, but also, as the poet makes explicit in the oration he wrote for the ceremony, a signal for “the impending reformation of an exhausted culture by means of a return to the spirit of the ancient” (Regn –​Huss 2009: 86). Petrarch’s reflections on fame and the legacy of classical tradition prompted the inscription of poetry (his own, as well as his forebears’ and contemporaries’) within the wider structure of human history. The memorial function of history was also a never-​ending struggle against time, as testified in the development of the Triumphi. At the same time, prophecy becomes for him another word for statement of purpose, as shown by the prophetic ending of the Africa. As late-​medieval English literature laboured to find its identity, in linguistic and cultural terms, the legacy of Petrarch proved essential, durable, and complex. From one generation of poets to the next, from Chaucer to Lydgate to Wyatt and Surrey, the Petrarchan texts drawn upon and the reactions they generated changed, sometimes radically, providing a singular instance of translatio

13 ‘Greedy Time wins and destroys all; it is called Fame, and it is a second death; and against it, as against the first, there is no remedy’. 14 ‘And after the books have died, you will also disappear; thus the third death awaits you’.

The Construction of the European Intellectual

95

studii: translations, rewritings, and parodies from Petrarch chart the passage of English writing from the late Middle Ages to the Renaissance. At the same time, Petrarch’s enfranchisement from the oppressive classical inheritance offers Chaucer and the following generation of poets a model for the building of a national literary canon, accompanying the emergence of English as the language of the nation. The poetic coronation (in keeping with the example of Albertino Mussato, the first laureate poet of Italy) honoured Petrarch as poeta et historicus (Petrina 2010). It is this sense of history as expressed by poetry that makes sense of him as a model for English poetry. If one of the texts on which this honour rests is Africa, the other is De viris illustribus, a triumphal procession that would find an enthusiastic following in literary and pictorial terms in England: John Lydgate’s Fall of Princes follows this model in being not simply a desolate catalogue of irrecoverably past glory, but an exhortation for the future. The idea of a “Renaissance project” may be linked to the appropriation of the notion of translatio studii: if Dante proposes a translatio imperii in Paradiso VI, and Petrarch appropriates this idea in his moving the site of the laureation between Rome and Paris,15 then the English poets, Chaucer in particular, may adumbrate a further translatio. Petrarch’s translation is a move back to classical culture: the figure is employed by Chaucer to suggest a move forward to the rarefied atmosphere of the north. Through history and geography, Petrarch looks for the collocation of man in the universe (woman obviously is still out of the equation). Such a role as that assumed by Petrarch asks us to reconsider the traditional division between Middle Ages and Renaissance: “the idea that the Renaissance discovered ‘the past as past’ [...] posits as a corollary vision the Middle Ages as a period of ‘diachronic innocence’ ” (Summit 2000: 213; see also Petrina 2019). This traditional division can be challenged also with reference to Chaucer: “Chaucer expresses a scepticism about periodization which is at least as modern as it is medieval. Rather than simply rejecting the new, he creates an authorial position that can consider the relationship between the Middle Ages and the Renaissance from both sides of the threshold at once, thereby calling into question the status of the threshold” (Stillinger 2009: 225). Petrarch asked his readers and followers to abandon this innocence, and his reception in Europe illustrates this decision. We may consider, for instance, a passage in Joachim du Bellay’s Deffence et Illustration de la Langue Françoyse:

15 “If Petrarch mentions in his coronation speech that the laurel crown was offered him not only by the city of Rome, but also by the University of Paris, it is in order to convey a programmatic message” (Regn –​Huss 2009: 87).

96

Alessandra Petrina qu’on me lyse un Demosthene et Homere Latins, un Ciceron et Vergile Françoys, pour voir s’ilz vous engendreront telles affections, voyre ainsi qu’un Prothée vous transformeront en diverses sortes, comme vous sentez, lysant ces aucteurs en leurs langues. Il vous semblera passer de l’ardente montaigne d’Aethne sur le froid sommet du Caucase. Et ce que je dy des langues Latine et Greque se doit reciproquement dire de tous les vulgaires, dont j’allegueray seulement un Petrarque, du quel j’ose bien dire, que si Homere et Virgile renaissans avoint enterpris de le traduyre, ilz ne le pouroint rendre avecques la mesme grace et nayfveté qu’il est en son vulgaire Toscan16 (Helgerson 2006: 334–​335).

Du Bellay is in fact asking us to ignore the classical/​vernacular dichotomy and acquire a new sense of time. Petrarch is posited not only as a supreme example of vernacular poetry, but also as the linchpin between the present and the past. Since the pivotal figure to trace the development of Petrarch’s influence in England is Chaucer, we are immediately faced with a problem. We have no clear sense of the extent of Chaucer’s knowledge of Petrarch: while normally we adhere to the comfortable persuasion that the European Middle Ages were familiar with the Latin Petrarch (Mann 1980), and the Renaissance with the Italian one, in the case of Chaucer we cannot conform to this principle. In two cases are we reasonably sure that Chaucer was translating from Petrarch –​in the already mentioned Canticus Troili, and in the Clerk’s Prologue and Tale, which also famously contains an explicit reference to the Italian poet.17 This does not mean that Sonnet 132 and the Historia Griseldis were the only Petrarchan writings Chaucer knew, but it does indicate that he dealt with a Petrarchan corpus that was substantially different from the one normally envisaged by medieval poets, and that the Clerk’s Tale presents a moment of conscious and explicit appropriation, in which the source is recognised with as much openness as if Petrach was a classical auctoritas. It also marks a conscious assumption of Chaucer’s role as translator, the same role for which he is famously celebrated by Eustace Deschamps (de Queux de Saint-​Hilaire –​Raynaud 1880: 138–​140). Beside translating the 16 ‘Just read a Latin Demosthenes and Homer, a French Cicero and Virgil to see if they will beget such emotions in you –​will, indeed, transform you like a Proteus into differing kinds –​as you feel reading those authors in their own languages. Going from the original to the translation, you will seem to pass from the burning mountain of Etna to the cold summit of the Caucasus. And what I say of the Latin and Greek languages can be equally said of all the vulgar tongues, of which I will cite only Petrarch, of whom I dare say that if a reborn Homer and Virgil undertook to translate him, they could not render him with the same grace and freshness that he has in his native Tuscan’. 17 “No previous tale, and indeed no later tale, is attributed by its teller to a single named author” (Stillinger 2009: 223).

The Construction of the European Intellectual

97

story, Chaucer imitates Petrarch’s stance as a traveller and explorer of books, and through the explicit act of translation expresses his awareness of his role: the Clerk does not excise Petrarch’s preamble. About to recount a tale by an author he venerates, he instead seems to split in two: one moment he is the deferential translator-​ copyist who will mirror in English the words he says he committed to memory at Padua; the next he is the assertive translator-​editor, ready not only to interpret but to trim the text as well. This double posture as humble pupil and bold redactor, which forecasts the Clerk’s identification with both Griselda and Walter in the tale, effectively translates Petrarch’s landscape into ethopoesis; the Clerk’s impulse to affirm his independence from authority and submit himself ungrudgingly to it becomes the principle that organizes all he says and does. (Ginsberg 2009: 159)

This passage from the translator-​copyist to the translator-​editor marks also the assertion of Middle English literature as a literature of England. Chaucer’s vision takes new breadth and space when he can use Petrarch as his source. The lines in which he celebrates Petrarch are well known: I wol yow telle a tale which that I Lerned at Padowe of a worthy clerk, As preved by his wordes and his werk. He is now deed and nayled in his cheste; I prey to God so yeve his soule reste! Fraunceys Petrak, the lauriat poete, Highte this clerk, whos rethorike sweete Enlumyned al Ytaille of poetrie, As Lynyan dide of philosophie, Or lawe, or oother art particuler; But Deeth, that wol nat suffre us dwellen heer, But as it were a twynklyng of an ye, Hem bothe hath slayn, and alle shul we dye. (IV. 26–​38)

But equally important are the following lines, in which the landscape expands, under Petrarch’s influence: But forth to tellen of this worthy man That taughte me this tale, as I bigan, I seye that first with heigh stile he enditeth, Er he the body of his tale writeth, A prohemye, in the which discryveth he Pemond and of Saluces the contree, And speketh of Apennyn, the hilles hye, That been the boundes of West Lumbardye, And of Mount Vesulus in special, Where as the Poo out of a welle smal

98

Alessandra Petrina Taketh his firste spryngyng and his sours, That estward ay encresseth in his cours To Emele-​ward, to Ferrare, and Venyse; The which a long thyng were to devyse. (IV. 39–​51)

A similar encapsulation of poetic authority within a geographical space is proposed by William Shakespeare (or John Fletcher) in the Prologue of Two Noble Kinsmen, when he praises Chaucer: We pray our play may be so, for I am sure It has a noble breeder and a pure, A learned, and a poet never went More famous yet ’twixt Po and silver Trent. (Potter 1997: 137–​38)

As Helen Cooper has written, such a formulation “sweeps into Chaucer’s shadow every writer from Petrarch to the new poets of the English Midlands” (Cooper 2006: 30). It is as if Shakespeare were picking up Chaucer’s suggestion and setting him in the position Chaucer himself had assigned to Petrarch. The Clerk’s Tale, with its amazing prologue, announces the truly European poet: geography is now a map seen from above, no longer a tortuous exploration. The prophecy at the end of Africa on dispersing darkness is fulfilled by Petrarch himself according to Chaucer, if the former has, in the Clerk’s words, enlumyned Italy (Stillinger 2009: 225–​26). By implication, successive generations of English poets will carry the illumination farther. If Chaucer offers us the first, astonishing intuition of what the Petrarchan legacy means to late medieval England, John Lydgate codifies it and makes it explicit. In order to trace wide-​ranging intellectual attitudes or changes in the late Middle Ages in England one should always go to the great encyclopaedic repository of Middle English writing, that is, John Lydgate’s works. Petrarch is mentioned in book 3 of the Troy Book (written for Henry, Prince of Wales, between 1412 and 1420), in the section concluding the long recapitulation of the Troilus and Criseyde’s story; Lydgate acknowledges Chaucer as the main poet who told this story, and this prompts a praise of the English poet, who deserves “the laurer of oure englishe tonge” (III.4246) for adorning with such eloquence his language, Right as whilom by ful highe sentence, Perpetuelly for a memorial, Of Columpna by the cardynal To Petrak Fraunceis was yoven in Ytaille (III.4248–​51)18

18 All quotations from the Troy Book are taken from Bergen (1906).

The Construction of the European Intellectual

99

The reference is very short and might be puzzling, especially as concerns the allusion to cardinal Colonna –​one might even suspect that Lydgate was confusing this Cardinal, Petrarch’s patron and friend, with the writer Guido delle Colonne, mentioned at the end of the Prologue (l. 360) as Lydgate’s main source. If this poem was dedicated to the future King Henry V, then the dedicatee was supposed to pick up the allusion immediately, to the point of seeing himself as a mirror image of the cardinal. On this parallel Lydgate constructs his vision of contemporary history: the relation between politics and poetry finds its affirmation in the poetic celebration of the great story of classical antiquity, the most political and most canonical of topics, the Troy war; by giving voice to this topic in a poem in English, Chaucer has given glory to the nation; implicitly, Chaucer’s disciple, Lydgate, will celebrate his royal patron and make him memorable as Petrarch had done for Cardinal Colonna. After all, Henry himself had asked him to compose the poem, as he notes in the Prologue (ll. 69–​80); Lydgate also makes it clear that the royal bidding was not simply the result of individual desire, but part of a wider political vision: Henry’s order is to translate the story of Troy into English, compiling the work after Guido’s original, By-​cause he wolde that to hyζe and lowe The noble story openly wer knowe In oure tonge, aboute in euery age, And y-​writen as wel in oure langage As in latyn and in frensche it is; That of the story θe trouthe we nat mys No more than doth eche other natioun. (Prologue, 111–​17)

The insistence upon the translation into “oure langage” reveals the poem’s political agenda: Derek Pearsall, commenting upon this passage, rightly notes that “the work is a status-​symbol, an attempt to define and consolidate the new status of English by tackling the greatest epic story of antiquity” (1970: 125). In this, Lydgate is simply faithful to his conventional role of courtly poet. But what follows is a passage exploring a recurring motif in Lydgate’s poetry, and possibly less interesting for his patron, that is, the importance of writing: For nere writers, al wer out of mynde, Nat story only, but of nature and kynde The trewe knowyng schulde haue gon to wrak, And from science oure wittes put a-​bak, Ne hadde oure elderis cerched out and souζt The sothefast pyth, the ympe it in oure thouζt, Of thinges passed, for-​dirked of her hewe, But thoruζ writyng θei be refresched newe,

100

Alessandra Petrina

Of oure auncetrys left to vs by-​hynde; To make a merour only to oure mynde, To seen eche thing trewly as it was, More briζt and clere θan in any glas. (Prologue, 159–​70)

The passage expresses an ongoing preoccupation with Lydgate, and we shall find similar meditations on the role of writing in The Fall of Princes. What I find striking here is that he has appropriated the Chaucerian image –​to make a mirror of our mind –​moving it from the courtly reflection on love to the humanist reflection on knowledge. We find yet another, fainter echo in Thomas Hoccleve’s Regement of Princes, in a passage in which he is celebrating Chaucer himself: Althogh his lyf be qweynt, the resemblance Of him hath in me so fressh lyflynesse That to putte othir men in remembrance Of his persone, I have heere his liknesse Do make, to this ende, in soothfastnesse, That they that han of him lost thoght and mynde By this peynture may ageyn him fynde. The ymages that in the chirches been Maken folk thynke on God and on his seintes Whan the ymages they beholde and seen, Where ofte unsighte of hem causith restreyntes Of thoghtes goode. Whan a thyng depeynt is Or entaillid, if men take of it heede, Thoght of the liknesse it wole in hem breede. (Blyth 1999: 186, ll. 4992–​5005)

If Chaucer takes from Petrarch the sense of space, Lydgate is fascinated by the Italian poet’s erudition. This has sometimes generated derision for the fifteenth-​ century poet. Yet, beyond a change of taste that makes his works very difficult to read for us, we recognize in Lydgate a never-​ending engagement with knowledge and writing, that echoes one of Petrarch’s traits: Petrarch’s socio-​cultural authority derived from his reputation as someone highly informed about matters not generally known, and that knowledge included esoteric and recondite geographical information deriving from both his first-​hand reconnaissance of places and from his investigations into the poetic and historical literature of the past. (Cachey 2003: 76)

In that incessant traveller of the mind and of Europe, Lydgate recognised his own aspirations, which expressed themselves in the form of endless catalogues, recognitions, retellings. Cachey continues by observing: The poet places his signature on various topoi of Roman imperial political geography, and prefigures, also in the hyperbole of the rhetorical gesture, the way in which early

The Construction of the European Intellectual

101

modern humanist self-​fashioners will attempt to inscribe themselves upon the new map of the world during the Renaissance, to support the illusion that they were the makers of their own worlds, or to simply assert their presence there. (Cachey 2003: 79)

In this adoption of the former poet’s intellectual attitude, Lydgate is the necessary linchpin between Middle Ages and Renaissance. In Lydgate’s most famous work, the incredibly articulate, encyclopaedic Fall of Princes, Petrarch is offered a much wider space. In the Prologue Lydgate laments that he has no Muse to turn to: Calliope will not listen to him, and Chaucer, his master and “cheeff poete of Breteyne”, is dead (ll. 239–​52). Chaucer is then inscribed in a series of great moral writers, from Seneca and Cicero to Petrarch and Boccaccio (ll. 253–​73). For each of these writers Lydgate also indicates some of their works. In the case of Petrarch, the list is re-​proposed, much amplified, later on; it includes most of his Latin production and makes no reference to either the Triumphi or the Canzoniere, and is set within a more general praise of writing, which even suggests a connection with Chaucer in its alluding to the “Hous of Fame” in which the poet is remembered: Writyng of old, with lettres aureat, Labour of poetis doth hihli magnefie, Record on Petrak, in Rome laureat, Which of too Fortunys wrot the remedie, Certeyn Ecloogis and his Cosmographie, And a gret conflict, which men may reede & see, Of his querellis withynne hymsilff secre. He wrot seuene Psalmys of gret repentaunce, And in his Affrik comendid Scipioun, And wrot a book of his ignoraunce Bi a maner of excusacioun, And sette a notable compilacioun Vpon the lyff[e]‌called solitarye, To which this world is froward and contrarie. And thus be writyng he gat hymsilff a name Perpetuelli to been in remembraunce, Set and registred in the Hous of Fame, And made Epistles of ful hih substaunce Callid Sine Titulo; & mor hymsilff tauaunce, Of famous women he wrot thexcellence, Gresilde preferryng for hir grete pacience. (IV. 102–​26)19

19 For Lydgate’s Fall of Princes the edition used is Bergen (1924–​27).

102

Alessandra Petrina

The works alluded to here are De remediis utriusque fortunae, Bucolicum carmen, Itinerarium breve de Ianua usque ad Ierusalem et Terram Sanctam (this is probably the work referred to as Cosmographie), Secretum, Psalmi penitentiales, Africa, De sui ipsius et multorum ignorantia, De vita solitaria, Sine nomine, and finally Petrarch’s Latin version of the story of Griselda: an impressively long list that includes most of Petrarch’s Latin output (including the Griselda translation), and obviously excludes the Canzoniere. Unlike Chaucer’s forays, Lydgate’s lists mirror the conventional medieval reception of the Italian poet and are themselves a monument to his achievement. A few stanzas later, Lydgate connects the idea of reading the works of poet as a way of finding our place in the universe: Men be writyng knowe the meracles Of blissid seyntes & of ther hoolynesse, Medecyne[s]‌, salue & eek obstacles Geyn mortal woundis and eueri gret seeknesse, Recreacioun and solace in distresse, Quiete in labour, in pouert pacience, And in richesse riht, trouthe and conscience. Shortnesse of lyff and foryetlnesse, The wit of man dul & ay slidyng, Necligence and froward idilnesse, –​ Echon stepmooder to science and konnyng, That I dar sey[e]‌n, nadde be writyng Onli ordeyned for our auauntages, Ded wer memorie & mynde of passid ages. (IV. 141–​53)

This concept might also be indebted to Petrarch, possibly via Chaucer, as the Italian poet developed the image, taken from Augustine’s Confessions, of reading and writing as a technique for describing the self (Stock 1995). Augustine writes in De utilitate credendi: “Cum legerem, per me ipse cognovi. Itane est?” (Stock 1995: 718).20 The idea is taken up by Petrarch who expresses himself through what he writes, but also by Chaucer, who, within the frame of a retraction, appends a full lists of his own works at the end of The Canterbury Tales. Lydgate imitates the gestures by defining himself through what he reads: The Fall of Princes presents a complete bibliography not only of Petrarch, but of Chaucer (I.274–​357), Virgil (IV.67–​91), and Ovid (IV.94–​105). Passages such as the ones quoted above show that the immense Fall of Princes is not simply the unstoppable progress of Joseph Ritson’s “driveling monk” (Ritson 1802: 87); it is a full-​length study on the 20 “When I read, it was I who gained knowledge through myself. Or was it?” (Stock 1995: 718).

The Construction of the European Intellectual

103

status of the poet, analysing the relation of modern poets to ancient ones, or the problems inherent in the transmission of poetry from one generation to another (as in the prologue to book VIII, in which Petrarch is seen visiting Boccaccio in a vision and giving him strength to continue his work), or the usefulness of comparing different versions of the same story; it is also, though in a somewhat garbled fashion, a modest proposal to re-​define the status of the writer. My last example is slightly later and is connected with Petrarch’s Secretum. The Secretum conforms to Petrarch’s public persona. It imitates and revises a classical genre (the Ciceronian dialogue); it stages a complicated debate between Christianity and classical culture; it combines an intimate, autobiographical idiom with pyrotechnic display of literary allusion; it meditates self-​consciously on its own practices of interpretation; it speaks to us of conflicts that seem, even after almost 700 years, recognizably modern. (Kahn 2015: 100)

The latter characteristic, perhaps in conjunction with the title,21 seems to be what has made the greatest impact on modern consciousness, and we tend to read the Secretum as if it was the key to Petrarch’s heart, as if Petrarch had unconscionably renounced his public persona to present us with his most intimate, “secret” one. But a striking instance of the late medieval English reception of this text suggests a different reading, linked rather to the display of the self and the almost theatrical qualities inherent in the form of the debate. British Library Additional 60577, also known as the Winchester Anthology, holds great interest for the cultural historian. Re-​discovered in 1979, after four centuries of oblivion, Additional 60577 is a mainly fifteenth-​century collection of verse and prose in English, French and Latin, all in the same hand, with some sixteenth-​century interpolation.22 As far as the original, fifteenth-​ century sections are concerned, the manuscript appears to have been compiled in St Swithun’s Priory, Winchester, in or around 1487. It seems to have been put together as a didactic volume, an anthology of mostly English and Latin texts, mainly of moral, religious or pedagogical import, connected with Winchester College, whose headmaster in the 1430s and 1440s was William Waynflete –​a 21 In the case of the title, the modern reader tends to miss the allusion to Augustine’s Confessions. Kahn (1985: 155) suggests the title rather highlights the function of this work as a reading of Augustine. 22 The manuscript was published in facsimile shortly after its rediscovery (Wilson –​ Fenlon 1981). Only a very small portion of the text has been edited so far, but very recently an edition of the translation of Secretum contained in the manuscript has appeared (Wilson –​Wakelin 2018).

104

Alessandra Petrina

name who appears and is celebrated in the manuscript. On fols 8r–​22v there is a text headed “Franciscus Petrarcha”, a Middle-​English translation of the Proem (fols 8r–​9v) and Book 1 of the Secretum. Once again, translation is a loose term. Petrarch’s original opens with the poet lost in a meditation that leads to his vision of Lady Truth: Attonito michi quidem et sepissime cogitanti qualiter in hanc vitam intrassem, qualiter ve forem egressurus, contigit nuper ut non, sicut egros animos solet, somnus opprimeret, sed anxium atque pervigilem mulier quedam inenarrabilis etatis et luminis, formaque non satis ab hominibus intellecta, incertum quibus viis adiisse videretur. Virginem tamen et habitus nuntiabat et facies. (Dotti 2000: 40)23

In the Middle English version, the initial state of meditation is expanded to a full-​ blown self-​presentation, which underlines the connection between Petrarch’s “Attonito michi quidem” and Boethius’ “Haec dum mecum tacitus”, the lines in De Consolatione Philosophiae (I p. 1, 1) introducing the apparition of Lady Philosophy (Stock 1995: 719–​20). I am soore astoned ‧ whan I remembre me How I entred thys lyff & how I schal oute agayne What ys thys world but wretchednes & aduersyte O ye lyff of man O vanyte & all vayne Where ys felycyte where. Hytt ys not here certayn Wee seche yt faste. & ouer ytt flethe awæye. Now god yt on ye crosse /​for our synne suffrede peyn Be ye oure guyde /​& brynge us in ye weye. Sumtyme I was a subtyle & a notable clerke And er I farther passe y wyll reherse myn name I am ye laureate poete called petrarc That in ytalye & Florence was of so grete Fame I appere to youre presence a processe to proclame Suffre me to saye /​I beseche yow myn entente And yff yt happe me to wrye thorowe dred or schame

23 ‘It happened recently. I was lost in thought, considering as I often do the way in which I came into this world and the way in which I must leave it; not overcome with sleep, as sick people often are, but wide awake with anxiety. Then I seemed to see a woman. she was from a time and of a splendour impossible to describe, and of a beauty which no mortal comprehends. By her clothes and general appearence she was a maiden’ (Nichols 2010: 5).

The Construction of the European Intellectual

105

Perdon me /​for in age to stumble yt ys convenient The matyer yt y purpose /​at thys tyme to trete Is callyd ye secrete conflycte of my conscyence …24

Petrarch’s revelation of truth, which developed from Boethius’ quest (in Boethius the Lady is Philosophy, in Petrarch it is Veritas), becomes in the Winchester Anthology, given the purpose of the codex, a form of teaching, introduced by the stanzas above, that to all intents and purposes announce the title and the author (Petrina 2013). If we posit, following the conventional reception of the Secretum in modern times, that it offers us a glimpse of the private Petrarch, intent on confession and self-​examination, then the additional lines by the English translator may be read simply as a form of advertisement, a proclamation of the well-​known auctoritas. But “ye laureate poete […] of so grete Fame” (ll. 11–​12) is also Petrarch in his own presentation. This version, which might have been meant for students’ recitation, and which occasionally alludes to a larger audience, preserves the very public persona Petrarch built for himself: the reception of the poet in England passes through the definition of the self by means of his writing, a man whose fame is inscribed within geographical and cultural boundaries, “in ytalye & Florence”. In his pivotal studies on Petrarch, Ernst Wilkins (1951) borrowed from the Metrical Epistle a memorable title for the poet: peregrinus ubique, a pilgrim everywhere.25 The metaphors of pilgrimage for reading and the pursuit of knowledge, the exploration and the travels of knowledge, often recur in Petrarch’s works. It is especially appropriate for Petrarch to have made his entrance in English literature thanks to that indefatigable scholarly pilgrim, Geoffrey Chaucer in his disguise as the Clerk of Oxford. Petrarch’s pilgrimages in England do not end here: roughly forty years after the Winchester Anthology was compiled, Thomas Wyatt came back from his Italian travels, and a new Petrarchism began in English poetry, not only through the partial translations of the Canzoniere and the fuller translations of the Triumphi, but also thought the re-​discovery and translation of De remediis (Twyne 1579). This paper has been re-​tracing the influence of Petrarch’s works in courtly, aristocratic poetry in England, leading to the early-​sixteenth-​century flourishing of northern humanism, with which it is directly connected. Such an exploration clashes not only against the conventional

24 London, British Library, Additional MS 60544, fol. 8r. I have silently expanded breviographs. 25 The tag comes from Metrical Epistle 3.19.16 to Barbato da Sulmona. Contini (1975: xv) uses the phrase “irrequieto turista” in his “Preliminari sulla lingua di Petrarca”. See Cachey (2003: 74).

106

Alessandra Petrina

division between a medieval, Latin Petrarch, and an early modern, Italian one, but also against the idea of a divide between the late Middle Ages and the early modern period in English literature: suffice it to think not only of the canonical status Chaucer would acquire in the Henrician era, but also of the important role Lydgate’s Fall of Princes played in the sixteenth century, or of the relevance of education, and of educational compilations, in the development of northern humanism. As a proto-​humanist, Petrarch longed to overcome the confines of his own time, looking back at classical past and forward at posterity; the early stages of his reception in England appear to follow the same pattern, and to challenge a standard periodization of literary ages.

References Ascoli, Albert Russell –​Unn Falkeid (eds.) 2015. The Cambridge Companion to Petrarch. Cambridge: Cambridge University Press. Andreoni Fontecedro, Emanuela. 2011. “Somnium Scipionis: emulazione nell’Africa del Petrarca e input dei dream poems di Chaucer”. Italica 88: 335–​352. Ariani, Marco (ed.). 1988. Francesco Petrarca. Triumphi. Milano: Mursia. Benson, Larry D. (ed.) 1988. The Riverside Chaucer. Oxford: Oxford University Press. Bergen, Henry (ed.). 1906. Lydgate’s Troy Book. (EETS ES 97). London: Kegan Paul, Trench, Trubner. Bergen, Henry (ed.). 1924–​27. Lydgate’s Fall of Princes. (EETS ES 121.124). London: Oxford University Press. Bernardo, Aldo S. (ed.). 1980. Francesco Petrarca Citizen of the World. Padova: Antenore. Blyth, Charles R. (ed.). 1999. Thomas Hoccleve. The Regiment of Princes. Kalamazoo: Medieval Institute Publications. Boswell, Jackson Campbell –​Gordon McMurry Braden. 2012. Petrarch’s English Laurels, 1475–​1700. Farnham: Ashgate. Cachey, Theodore J. jr. 2003. “Petrarchan Cartographic Writing”. In: Gersh, Stephen –​Bert Roest (eds.): 73–​91. Collette, Carolyn P. 2018. “Richard de Bury, Petrarch and Avignon”. In: Fulton, Helen –​Michele Campopiano (eds.): 40–​51. Contini, Gianfranco (ed.) 1975. Canzoniere. Torino: Einaudi. Cooper, Helen. 2006. Shakespeare and the Middle Ages. Inaugural Lecture Delivered 29 April 2005. Cambridge: Cambridge University Press.

The Construction of the European Intellectual

107

D’Agata D’Ottavi, Stefania. 2011. “Guido Cavalcanti’s Theory of Love and Chaucer’s Troilus and Criseyde”. Textus 24: 427–​448. de Bruyne, Edgar. 1946. Études d’Esthétique Médiévale. Brugge: Rijksuniversiteit te Gent. de Queux de Saint-​Hilaire, A.H.E. –​G. Raynaud (eds.) 1880. Oeuvres Complètes de Eustache Deschamps. Vol. 2. Paris: Didot. Dotti, Ugo (ed.). 1991. Francesco Petrarca. Le familiari. Libro secondo. Roma: Archivio Guido Izzi. Dotti, Ugo (ed.). 1994. Francesco Petrarca. Le familiari. Libro terzo. Roma: Archivio Guido Izzi. Dotti, Ugo (ed.). 2000. Francesco Petrarca. Il mio segreto. Milano: Rizzoli. Freccero, John. 1975. “The Fig Tree and the Laurel: Petrarch’s Poetics”. Diacritics 5: 34–​40. Fulton, Helen –​Michele Campopiano (eds.). 2018. Anglo-​Italian Cultural Relations in the Later Middle Ages. University of York: York Medieval Press. Galloway, Andrew –​R.F. Yeager (eds.). 2009. Through a Classical Eye. Transcultural and Transhistorical Visions in Medieval English, Italian, and Latin Literature in Honour of Winthrop Wetherbee. Toronto: University of Toronto Press. Gersh, Stephen –​Bert Roest (eds.). 2003. Medieval and Renaissance Humanism: Rhetoric, Representation and Reform. Leiden: Brill. Ginsberg, Warren. 2009. “From Simile to Prologue: Geography as Link in Dante, Petrarch, Chaucer”. In: Galloway, Andrew –​R.F. Yeager (eds.): 145–​164. Ginsberg, Warren. 2011. “Chaucer and Petrarch: ‘S’amor non è’ and the Canticus Troili”. Humanist Studies and the Digital Age 1: 121–​127. Helgerson, Richard (ed. and transl.). 2006. Joachim du Bellay. The Regrets with The Antiquities of Rome, Three Latin Elegies, and The Defense and Enrichment of the French Language. Philadelphia: University of Pennsylvania Press. Kahn, Victoria. 1985. “The Figure of the Reader in Petrarch’s Secretum”. PMLA 100: 154–​66. Kahn, Victoria. 2015. “The Defense of Poetry in the Secretum”. In: Ascoli, Albert Russell –​Unn Falkeid (eds.): 100–​110. Kaylor, Noel Harold Jr. 1993. “Boethian Resonance in Chaucer’s “Canticus Troili’ ”. The Chaucer Review 27: 219–​227. Lenoir, Rebecca (ed.). 2002. Pétrarque. L’Afrique : 1338–​1374, Grenoble : Millon. Mann, Nicholas. 1980. “Petrarch and Humanism: The Paradox of Posterity”. In: Bernardo, Aldo S. (ed.): 287–​299. Mann, Nicholas. 1984. Petrarch. Oxford: Oxford University Press.

108

Alessandra Petrina

Nichols, J.G. (transl.). 2010. Petrarch. Secretum. Richmond: Oneworld Classics. Pearsall, Derek. 1970. John Lydgate. London: Routledge & Kegan Paul. Petrina, Alessandra. 2010. “ ‘With his penne and langage laureate’: The Symbolic Significance of the Laurel Crown”. Studi Petrarcheschi 23: 161–​185. Petrina, Alessandra. 2013. “The Humanist Petrarch in Medieval and Early Modern England”. Journal of Anglo-​Italian Studies 12: 45–​62. Petrina, Alessandra. 2019. “All Petrarch’s Fault: The Idea of a Renaissance”. Memoria di Shakespeare 6: 145–​164. Potter, Lois (ed.). 1997. William Shakespeare. The Two Noble Kinsmen. Walton-​ on-​Thames: Thomas Nelson. Regn, Gerhard –​Bernhard Huss. 2009. “Petrarch’s Rome: The History of the Africa and the Renaissance Project”. Modern Language Notes 124: 86–​102. Ritson, Joseph. 1802. Bibliografia Poetica: A Catalogue of English Poets. London: Roworth. Rossiter, William T. 2010. Chaucer and Petrarch. Cambridge: D.S. Brewer. Stewart, H.F. –​E.K. Rand –​S.J. Tester (eds. and transl). 1973. Boethius. The Theological Tractates. Cambridge, Mass.: Harvard University Press. Stillinger, Thomas C. 2009. “New Science, Old Dance: The Clerk and the Wife of Bath at Philology”. In: Galloway, Andrew –​R.F. Yeager (eds.): 223–​38. Summit, Jennifer. 2000. “Topography as Historiography: Petrarch, Chaucer, and the Making of Medieval Rome”. Journal of Medieval and Early Modern Studies 30: 211–​46. Stock, Brian. 1995. “Reading, Writing and the Self: Petrarch and his Forerunners”, New Literary History 26: 717–​30. Twayne, Thomas (transl.) 1579. Phisicke against fortune, aswell prosperous, as aduerse conteyned in two books. London. Richard Watkyns. Wilkins, Ernst H. 1943. “The Coronation of Petrarch”. Speculum 18: 155–​97. Wilkins, Ernst H. 1951. The Making of the ‘Canzoniere’ and other Petrarchan Studies, Roma: Edizioni di Storia e Letteratura. Wilson, Edward –​Iain Fenlon (eds.) 1981. The Winchester Anthology. A Facsimile of British Library Additional Manuscript 60577 with an Introduction and List of Contents by Edward Wilson and an Account of the Music by Iain Fenlon. Cambridge: D.S. Brewer. Wilson, Edward –​Daniel Wakelin (eds). 2018. A Middle English Translation of Petrarch’s Secretum. (EETS OS 351). Oxford: Oxford University Press.

Part II:  Borrowing and Lexicon

Angelika Lutz

Changes of Political Rule and the Changing Use of OE Gærsum(a) ‘Treasure’ (< ON Gersemi) in Middle English Abstract: The late Old English attestations of the Norse loan gærsum(a) ‘treasure’ reflect historical facts resulting from Cnut’s rule over late Anglo-​Saxon England, just like the late Old English uses of lagu ‘law’ and eorl ‘regional leader’. Whereas law has become the most general legal term and earl part of the post-​Conquest hierarchy of otherwise largely Norman French loans, most other terms reflecting foreign rule borrowed from Old Norse were replaced with synonyms borrowed from Norman French. The uses of ME gersum(e) are shown to represent an alternative to both wholesale survival and abandonment: Although as a general term, the Norse loan was replaced with synonymous tresor borrowed from Norman French in the first half of the twelfth century, gersum(e) continued to be used with the same meaning until the late fifteenth century. These uses in various regions culturally related to the former Danelaw were restricted to fictional and religious texts. Keywords: loan-​word, Norman French, Old English, Old Norse, historical fiction

1. Introduction: Lexical Reflections of Norse Rule in Late Old English and Beyond Among the words borrowed from Old Norse that reflect the influence of Viking rule on the late Old English lexicon, two may be singled out as particularly remarkable: OE lagu ‘law’, which replaced the inherited synonym ǣw and has become the most general term referring to legal rule. And OE eorl ‘regional leader’, which replaced the inherited synonym ealdorman as a term referring to representatives of political and military rule directly below the king and later established itself as an integral part of the post-​Conquest hierarchy of terms referring to ranks otherwise consisting mostly of Norman French loans. Both Norse loans are known to have become part of the Old English lexicon before Cnut’s time, as regional terms with reference to the conditions in the so-​called Danelaw. But under Cnut’s rule, they both gained nation-​wide relevance and have retained it until today. The details of their lexical survival and semantic development have

112

Angelika Lutz

been the subject of various studies.1 Less attention has been devoted to the fact that the two loans survived as etymological isolates because most Norse loans referring to Viking rule in late Anglo-​Saxon England were soon after replaced with Norman French synonyms, as a result of the change to Norman rule (see Lutz 2012: section 2.2). The present paper focuses on a Norse loan whose Old English uses can likewise be associated with Cnut’s rule and its consequences but whose Middle English development represents an interesting alternative to both wholesale survival and total abandonment: The earliest use of the Norse loan gærsum(a) ‘treasure’ reflects historical facts related to Cnut’s rule over late Anglo-​Saxon England.2 From then on, the word was frequently used in historical contexts until the early Middle English period, when gersum(e) was replaced as a general term with synonymous tresor borrowed from Norman French. The use of the Norse loan became restricted to text types in which the word referred to treasures in less obviously real worlds than that of politics, i.e. to fictional and religious texts. In such texts, the Norse loan continued to be used until the late Middle Ages.

2. Uses of the Norse Loan for ‘Treasure’ in the Anglo-​Saxon Chronicle as Reflections of Political Influence The earliest uses of the Norse loan are attested in several manuscript versions of the Anglo-​Saxon Chronicle, which allows for a relatively precise dating and localisation of the evidence in this closely interrelated group of historical texts. Some of the Chronicle versions are clearly Old English, some are rightly listed as evidence for early Middle English.3 The first attestation of OE gærsum(a) ‘treasure, precious object(s)’ (< ON gersemi), namely in the annal for 1035 in MS C of the Chronicle, can be related to Cnut’s rule. Basically the same textual version is also

1 See OED s.vv. law and earl. For comprehensive assessments of the influence of Old Norse on the Old English lexicon see esp. Peters (1981), Townend (2002), and Pons-​ Sanz (2013), for the development of Norse loans referring to legal and political rule Fischer (1989) and Lutz (2012: section 2.2 and 2013: section 3). 2 For earlier discussions of this word, each with a different focus, see Lutz (2018) and Lutz (2019: section 2.8). 3 See DOE s.v. gærsum m./​n. and gærsuma f., MED s.v. gersum(e) and de Vries s.v. gersemi f. ‘kleinod, kostbarkeit’. For the complex textual relationship between the extant manuscript versions of the Chronicle see Keynes (2012) and the introductory chapters of the four later versions (C, D, E, F) referred to in note 4 below.

Changes of Political Rule and the Changing Use of OE

113

attested in MS D.4 It reports in some detail on the death of King Cnut and the conflict between his widow Ælfgifu /​Emma and her stepson Harold Harefoot over his father’s treasure, who claimed it as Cnut’s successor:5 (1) s.a. 1035 C(D). Her forðferde Cnut cing […] and hine man ferode þanon to Winceastre and hine þær bebyrigde. And Ælfgyfu seo hlæfdie sæt þa ðær binnan, and Harold […] let niman of hyre ealle þa betstan gærsuma, […] þe Cnut cing ahte. ‘In this year King Cnut died […] and they brought him from there to Winchester and buried him there. And Ælfgifu, the Queen, then stayed there. And Harold […] had all the best treasures taken from her, […] which King Cnut possessed.’

It is possible that this attestation of the Norse loan in the two eleventh-​century manuscripts C and D reporting on Cnut’s death represents its first written use as an Old English loanword. But it is likely that the word had become current in oral exchanges during Cnut’s lifetime among members of his royal circle at Winchester not only in his native Old Norse but also in Old English. At any rate, from then on, the loanword gærsum(a) was used in Chronicle entries for almost a century, frequently in contexts in which it referred to royal power and not just to the monetary value of a treasure. The frequently attested plural forms, here that of a strong neuter, allow for both plural and collective interpretations.6 In earlier Old English mentions of treasures, likewise attested in the Chronicle, the meaning ‘treasure’ is expressed with the inherited words māðum and sceatt, where the latter word mostly refers to the monetary value of a treasure.7

4 For the text in MS C, which may be linked with Abingdon, see O’Brien O’Keeffe (2001: 105), for the version in D, probably from Worcester, see Cubbin (1996: 65), for its translation Whitelock (1979: 256–​257). The other two late versions of the Chronicle, MSS E and F, for this period both associated with Canterbury (see Irvine 2004: lxiv–​ lxxv and Baker 2000: xxviii–​xxxix), do not mention this conflict over Cnut’s treasure in the royal family. 5 For Cnut’s burial in the Old Minster, much later also that of his second wife Ælfgifu /​Emma, and the post-​Conquest transfer of their bodies to the Norman cathedral see Biddle –​Kjølbye-​Biddle (2016: 212–​228). 6 Cf. DOE s.v. I.a and I.b. 7 See Clark Hall s.v. māðum ‘treasure, object of value, jewel, ornament, gift’ and sceatt ‘property, treasure, coin, money, wealth’, TOE (I: 640, 15.01.03), and Orel s.vv. *maiþmaz and *skattaz. The use of māðum is known from King Alfred’s famous preface to his Pastoral Care: “hu ða ciricean giond eall Angelcynn stodon maðma ond boca gefylda” (‘how the churches throughout England stood filled with treasures and books’) but is also attested in the Chronicle (see Pons-​Sanz 2013: 235–​236).

114

Angelika Lutz

The second attestation of the Norse loan, from MS D only, reports how Edward the Confessor took possession of the royal treasure from his mother Ælfgifu /​ Emma with the political and military support of his earls in 1043, shortly after his coronation:8 (2) s.a. 1043 D. Her wæs Eadward gehalgod to cynge æt Wincestre […] and þæs geres […] man gerædde þan cynge þæt he rad from Gleawecestre, […] and Leofwine eorl and Godwine eorl and Sigwarð eorl mid heora genge, to Wincestre on unwær on þa hlæfdian, and bereafdan hi æt eallon þan gærsaman þe heo ahte […] ‘In this year Edward was consecrated king at Winchester. […] And this year […] the king was advised to ride from Gloucester, […] together with Earl Leofric and Earl Godwine and Earl Siward and their retinue, to Winchester unexpectedly upon the lady, and they deprived her of all the treasures which she owned […]’

Many of the Chronicle attestations can be dated relatively closely to the dates provided by the annal numbers. This certainly applies to the altogether five occurrences in the late-​eleventh-​century MS D9 and at least partly to the altogether sixteen occurrences in MS E, which in its attested form is a twelfth-​ century text produced by two scribes in Peterborough. All sixteen occurrences in E are attested in the portion of the first scribe, who is responsible for the copy until 1121 and his continuation until 1131.10 An obvious exception to this are the two attestations from the bilingual MS F written at Christ Church (Canterbury) shortly after 1100, where the late F-​scribe employs the loan for a detailed local addition to the entry s.a. 995.11 These two post-​Conquest attestations of the word in F with reference to a much earlier event are nevertheless of variational interest insofar as they suggest that the Norse loan was current in early post-​Conquest times in Canterbury and not only in the two partly Danelaw-​related manuscripts D and E associated with Worcester and Peterborough, respectively.

8 For the text in D see Cubbin (1996: 66–​67), for its translation Douglas –​Greenaway (1981: 105). The parallel version for this date in MSS C and E does not speak of a treasure literally but of “gold and silver and things beyond description” (translation ibid.). For the situation after King Harthaknut’s death and for Ælfgifu /​Emma’s role as mother of Harthaknut (by Cnut) and Edward (by Ethelred) see esp. Keynes (1998: lxviii–​lxxviii). 9 In addition to the attestations s.aa. 1035 and 1043, the loan occurs in D s.aa. 1075 (2) and 1078 (see Cubbin 1996: 67, 86, 88). 10 For the sixteen attestations s.aa. 1047, 1070 (5), 1076, 1086 (3), 1090 (2), 1095, 1122, and 1128 (2) see Irvine (2004: xviii–​xxii, 80–​130). 11 Namely with reference to the dealings of Archbishop Ælfric of Canterbury with the pope; see Baker (2000: 91–​92) and Whitelock (1979: 236 note 3).

Changes of Political Rule and the Changing Use of OE

115

The two latest Chronicle attestations of the loan nicely illustrate the political importance of having a gersum(e) also for a Norman king, in an entry s.a. 1128 on the visit paid by Hugo of Jerusalem to King Henry, the Conqueror’s son:12 (3) s.a. 1128 E. […] And se kyng him underfeng mid micel wurðscipe and micele gersumes him geaf on gold and on siluere. And siððon he sende him to Englalande, and þær he wæs underfangen of ealle gode men, and ealle him geauen gersume and on Scotlande ealswa. ‘And the king received him with great honour and gave him [large] treasures consisting of gold and silver. And then he sent him to England where he was received by all good men, and they all gave him treasures and in Scotland also.’

Taken together, the four examples of the use of the Norse loan from the eleventh and early twelfth centuries discussed here seem to present a uniform picture insofar as they were all used to refer to valuables of the real world that played a role in the execution of power. This assumption is supported by the Chronicle evidence as a whole.13 And the fact that we have twelfth-​century attestations both from southern Canterbury and from Danelaw-​related Peterborough suggests that the noun could have been on the way to establishing itself as a general term in the entire country, similar to the Norse loans for ‘law’ and ‘regional leader’. However, the latest section of MS E, written by the scribe of the Second Continuation after the middle of the twelfth century, betrays that this Norse loan was replaced with the Norman French loan tresor shortly afterwards even in this Danelaw-​related Chronicle version:14 (4) s.a. 1137 E. Đis gære for þe king Stephne ofer sæ to Normandi and ther wes underfangen, forþi ðat hi uuenden ðat he sculde ben alsuic alse the eom wes, and for he hadde get his tresor. ‘This year king Stephen went overseas to Normandy and was received there because they expected that he would be just as his uncle had been, and because he still had his treasure.’

12 For the text see Irvine (2004: 130), for the translation Douglas –​Greenaway (1981: 205). For the First Continuation, which comprises the entries s.aa. 1122–​1131, see Clark (1970: xv–​xxv) and Skaffari (2002: 237–​244). 13 Sara Pons-​Sanz, who likewise notes the frequent use of the Norse loan in the Chronicle, attributes it to the fact “that most English territories would have been affected by the Scandinavians’ taste for riches and precious objects and their willingness to take them away, by force if necessary” (Pons-​Sanz 2013: 236). 14 For the text see Irvine (2004: 134), for the translation Douglas –​Greenaway (1981: 210). The Second Continuation comprises the entries s.aa. 1132–​1154 (cf. Clark 1970: xxv–​ xxviii and Irvine 2004: xxii–​xxiii). For the influences of Old Norse and Norman French attested in this last section of the Peterborough Chronicle see Skaffari (2002: 237–​244).

116

Angelika Lutz

Just like the three attestations of the Norse loan in the earlier annals cited above, this very first attestation of the Norman French loan tresor illustrates the political relevance of a treasure for a king. And since England was not subjected to any other foreign rule after that, the Norman French loan became the general term for ‘treasure’ and has remained in use as such until today.15

3. Uses of ME Gersum(e) in Historical Fiction Despite the early-​ twelfth-​ century replacement of the Norse loan with the Norman French loan as the general term for ‘treasure’ in the Chronicle, the noun gersum(e) remained in use with the same meaning until the late fifteenth century. The MED distinguishes three types of uses: 1a. “A treasured object, valuable possession”; 1b. coll. “Treasure, valuables”; and 1c. In alliteration: gold and gersum. And it cites several examples from the First Continuation of the Peterborough Chronicle as the earliest Middle English attestations of 1a and 1b. Closely related to the Chronicle with regard to subject matter and lexical choices are a number of Middle English texts that may be characterized collectively as examples of historical fiction. The distinction between historical narrative texts with a straightforward historiographical function and historical fiction remains difficult even beyond medieval times, despite the lexical distinction between history and story that began to be made in England from very late medieval times onwards.16 The compilators of the Helsinki Corpus therefore saw good reasons for listing both types of texts under “History” for much of the Middle English period, e.g. the two Continuations of the Peterborough Chronicle and Laʒamon’s Brut under ME I (1150–​1250). Middle English texts of the latter type with attestations of gersum(e) include the early Middle English Brut and several much later texts of the so-​called Alliterative Revival that are linked with the Brut (a) with regard to 15 See MED s.v. tresour and OED s.v. treasure. For particularly obvious lexical reflections of Norman rule in post-​Conquest English see also the rich attestation of the nouns castel, prisun and peis in the late entries of the Peterborough Chronicle (cf. Skaffari 2009: 154–​157) and the replacement of the entire word family of OE æþele with the word family of noble (cf. Vennemann 2012). 16 Cf. OED s.vv. This is also reflected in Sir Robert Cotton’s press marks of numerous Old and Middle English texts, which have survived in the British Library pressmarks e.g. of five of the seven manuscripts containing the Anglo-​Saxon Chronicle (Tiberius A.vi, Tiberius B.i, Tiberius B.iv, Domitian A.viii, Otho B.xi), of Beowulf (Vitellius A.xv), Laʒamon’s Brut (Caligula A.ix, Otho C.xiii), and Sir Gawain and the Green Knight (Nero A.x), to name a few. They betray an intense interest of early modern scholars in both types of text.

Changes of Political Rule and the Changing Use of OE

117

content, i.e. historical bent and Arthurian subject matter, and (b) formally, by their preference for alliteration to rhyme. But they also include Cursor Mundi, which differs in content and employs end rhyme. Details about the three types of historical fiction and their uses of gersum(e) will be provided below, in sections 3.1 to 3.3.

3.1. Uses of ME Gersum(e) in Laʒamon’s Brut Laʒamon’s Brut consists of ca 16,000 long-​lines with alliteration and rhyme and was composed around 1200.17 The text exists in the form of a longer and lexically more conservative manuscript version (Cotton Caligula) and a considerably shorter, lexically more progressive manuscript version (Cotton Otho), both probably from the later thirteenth century. Laʒamon’s text, which may be considered a late reflection of Old English heroic poetry, deals with the pre-​history of Britain and Anglo-​Saxon England.18 It is the earliest and most substantial Middle English fictional text with attestations of the Norse loan gersum(e). Laʒamon characterizes himself as a priest from near Worcester, and we know that Worcester had close political and cultural ties with York over a long period of time stretching across the Norman Conquest.19 The following examples of gersum(e) from the Brut show that Laʒamon employed the loan in various ways though mostly for alliteration within the long-​line. These examples show that the distinction in the MED between (1b) collective use and (1c) in alliteration with gold as a more or less fixed combination is not always helpful for studying Laʒamon’s varying uses of the word.20 In example (5a), the Caligula version alliterates gersum with gold in the same long-​ line but also links gold with silver as parts of the second half-​line and then, in the following line, varies gersume with the older, inherited synonym maðmas:

17 For the Brut see Le Saux (1989: ch. 1), Allen (1992: xiii–​xxxiv), Roberts (1994), Barron –​ Weinberg (1995: ix–​xx), and Elsweiler (2011: 1–​5). 18 Based on Geoffrey of Monmouth’s Historia Regum Britanniae and Wace’s Roman de Brut (see Le Saux 1989: chs. 4 and 5, Allen 1992: xiii–​xix, and Barron –​Weinberg 1995: ix–​xx). 19 For this particular cultural contact and its linguistic effects see Dance (2003: ch. 1, esp. pp. 33–​35). 20 For the text of the Brut see Brook –​Leslie (1963, 1978). My discussion of the examples of the noun in the Brut is not based on a full list of attestations but largely on those provided in the MED s.v.

118

Angelika Lutz

(5) a. […] ʒif he me ʒefeð gersume; gold & seoluer.    alle his maðmas; wið þon þa he mote libben (Brut, Cal. 444–​45)   ‘[…] if he gives me a treasure, gold and silver, /​all his riches; with that he may live’

The Otho version employs the loan as part of the alliterating pair gold and garisom in the second half-​line and in the following line continues the alliteration pattern of the previous line by way of replacing maðmas with roughly synonymous godes: (5) b. […] ʒif he vs ʒiue wolle; gold and garisom.   and alle his godes wid þan þe he mote libbe (Brut, Ot. 444–​45)   […] if he is willing to give us gold and riches /​and all his possessions; with that he may live’

In the MED, the Caligula reading is therefore listed under (1b), the Otho reading under (1c). In the following pair of examples, Caligula exhibits alliteration of gersum(e) (second half-​line) with both Godlac and the second element of the compound scip-​gumen (first half-​line). Otho, by replacing -​gumen with -​men, reduces the number of alliterating elements in this line, which is in harmony with Otho’s generally much less frequent use of the noun gume:21 (6) a. Godlac sloh þa scip-​gumen; & alle heo nom þat garsume (Brut, Cal. 2276) (6) b. Gutlac sloh þe sipmen; & nam al þe garisom (Brut, Ot. 2276)   ‘Godlac killed the shipmen; and took all the treasure’

The next example is only attested in Caligula, due to Otho’s overall shortening tendencies. Here, the loan in the second half-​line alliterates with the stressed syllable of the adverb in the first half-​line: (7)

He somnede to-​gædere; gærsumme muchele (Brut, Cal. 3271) ‘He carried together a very great treasure’

The following example, where gersum(e) in the second half-​line alliterates with gumen in the first half-​line, likewise occurs only in Caligula because Otho has been shortened: (8)

Vortiger ʒef þissen gumen; swiðe muchele gærsume (Brut, Cal. 6713) ‘Vortiger gave these men very great wealth’

In the following example, the longer Caligula version comes close to using gold and gersum(e) in the form of an alliterating formula. In the shortened Otho text, 21 See Elsweiler (2011: 105), who lists altogether 75 attestations of gume in Caligula and only 36 in Otho.

Changes of Political Rule and the Changing Use of OE

119

the formula is abandoned and garisome of the first line alliterates with ʒiftes gode of the second line: (9) a. […] & auerælcne cnihtes sune; þat his main wolde cuðen.   and ælcne wiþerfulne mon; þe god wolde biwinnen.   þat him to scolden comen; for golde and gæirsume (Brut, Cal. 5272–​74)   ‘[…] & every knight’s son that wanted to test his strength, /​and every wicked  man who wanted to gain wealth, /​that they should come to him for gold and treasure’ (9) b. […] and alle þe cnihtenes sones; þat wolde habbe garisome.   þat hii to him solde come; and habbe ʒiftes gode (Brut, Ot. 5272 –​74)  ‘[…] and all the knights’ sons that wanted to have treasure, /​that they should come to him and receive good presents’

This contrasts with examples (5a) and (5b) above, where only Otho alliterates gersum(e) with gold within the half-​line. Thus, taken together, the examples from the Brut exhibit a considerable variety of uses of the Norse loan for ‘treasure’. The word is always linked by alliteration though not always within the long-​line. The contexts in which the Norse loan was used in this piece of historical fiction are similar to the contexts of the Chronicle prose of the twelfth century.

3.2. Uses of ME Gersum(e) in Texts of the Alliterative Revival These late Middle English texts that are similar to the much earlier Brut with regard to their Arthurian subject matter and their preference for alliteration to rhyme were composed in areas concentrating on but not restricted to the former Danelaw.22 Since the Helsinki Corpus does not use these texts but Malory’s non-​ alliterating Morte Darthur (under “Romance”), it remains unclear whether these texts would have been counted as “History” or as “Romance”. The following example of the late Middle English use of gersum(e) is from Awntyrs off Arthur, a relatively short text composed ca 1400–​1430 in the North-​East Midlands. This text has survived in a manuscript dated to the late fifteenth century.23 The MED s.v. 1a lists it as an example of plural use as a count noun, the first after three

22 For an understanding of the texts that belong to the Alliterative Revival and for their complex cultural background see especially Pearsall (1981: 14–​17), Lawton (1982), Pearsall (1982), and Hanna (2004). For later medieval uses of the Brut see McNelis III (1994). 23 For the text, with glosses, see Hanna (1974: 96), for the dating and localisation of the text and the four manuscripts ibid. pp. 1–​11.

120

Angelika Lutz

attestations from the Peterborough Chronicle. Here, the loan is linked with gyftis in an alliterative nominal pair: (10) There he wedded his wife, wlonkest I mene Withe gyftis and gersones, Sir Galeron the gay (Awntyrs Arth. 696–​97) ‘There he wedded his wife, the richest I believe, /​with regard to gifts and treasures, Sir Gallerone the splendid one’

The last example of this use cited in the MED is from the alliterative Morte Arthure dated to ca 1440 and localized to the North of England. Here, the loan likewise alliterates but not as part of an alliterative nominal pair:24 (11) þou sall haue gersoms full grett þat gayne sall þe euere (Morte Arth. (1) 165) ‘You will have great treasures that will always serve you’

Otherwise, the MED cites examples from the Alliterative Revival as alliterating pairs with gold (s.v. 1c). In such combinations, the loan is attested in several poems of the late fourteenth and fifteenth centuries, in Sir Firumbras, Morte Arthure, and Awntyrs off Arthur. Sir Firumbras, dated to ca 1380, survives in a copy from the far south-​western region:25 (12) Of þat gold & of þat geryzoun þat he fond on the tour & on þe toun (Sir Firumb. (1) 5693) ‘Of the gold and treasure that he found in the tower and the town’

In the following example from Awntyrs off Arthur, the alliterating pair forms part of a solemn, possibly formulaic declaration in direct speech:26 (13) Here I gif Sir Gawayne, with gersone and golde,  Al þe Glamergane londe with greves so grene (Awntyrs Arth. (Dc) 664–​65) ‘Here I give Sir Gawain, with treasure and gold, /​all the land of Glamorgan with groves so green’

Thus, the Middle English evidence for the Norse loan from historical fiction demonstrates that the word remained in use until very late. But whereas its use in the early Middle English Brut shows considerable variation (see 3.1. above), its use in the late Middle English texts of the Alliterative Revival seems to have become restricted to few, largely fixed nominal pairs.

24 For the text see Benson (1994), but with the manuscript spelling from the MED s.v. gersum(e) 1a. 25 See Herrtage (1879: 177) and LALME I (145). 26 See Hanna (1974: 6–​11, 94).

Changes of Political Rule and the Changing Use of OE

121

3.3. Uses of ME Gersum(e) in Cursor Mundi Cursor Mundi deals with the history of the Christian world in ca 30,000 short couplets and thus differs from the Brut and the Arthurian romances of the Alliterative Revival both in subject matter and form.27 This long text, composed by a clearly Northern author, has survived in nine manuscripts.28 MS C (= London, BL, Cott. Vespasian A.iii) can be dated to ca 1340 and is one out of three Northern manuscripts. Only this manuscript contains both of the altogether two attestations of the loan gersum(e), both of them for collective use (MED s.v. 1b). The first alliterates with rhyming gifte: (14) If theif na gersum has ne gifte, þat he mai yeild again his thift, He sal be saald […] (CM 6753–​55) ‘If a thief has neither a treasure nor a present, /​that he may make amends for his theft, /​he shall be sold […]’

In MS F (= Oxford, Bodl. Fairfax 14) from the late fourteenth century, which is associated with Lancaster, gersum(e) has been replaced with the alliterating equivalent gode, whereas MS G (= Göttingen University theol. 107) from the second half of the fourteenth century has the non-​alliterating equivalent fine, a Norman French loan.29 The two manuscripts of the Southern version, T and H, likewise read fine. The second attestation of gersum(e) in Cursor Mundi occurs in two of the three Northern manuscripts, C and G:30 (15) Bot stedfast hope and trout right, And ert clene and eie Oþir gersum ask i nan (CM 25469–​71) ‘But steadfast hope and faithful trust, /​And pure heart and eye, /​I ask for no other reward’

In MS F, gersum(e) has been replaced with the French loan couette ‘desire, wish’, the Southern versions of this section of the text do not survive. This suggests 27 The compilers of the Helsinki Corpus list this fourteenth-​century text under “History”, just like the Peterborough Chronicle and the Brut. 28 See Morris’s edition (1961–​1966) for the leftmost of four parallel manuscript versions. All nine manuscripts are described in Horrall (1978: 13–​23). 29 For the manuscript texts see Morris (1961–​1966: 390–​391) and Horrall (1978: 256), for the widespread legal use of the French loan fine in later Middle English see MED s.v. 5–​11. 30 For the texts see Morris (1961–​1966: 450–​451).

122

Angelika Lutz

that by the fourteenth century the loan had already become rare in the poet’s Northern dialect and that the scribes of the Northern recension seem to have felt increasingly uneasy to use the word, possibly because they were no longer sure of its meaning. The clearly less frequent use of gersum(e) in Cursor Mundi than in the earlier Brut and the later Arthurian romances of the Alliterative Revival may be due to its different subject matter and to the use of end rhyme instead of alliteration.

4. Uses of ME Gersum(e) in Religious Texts The attestations of the Norse loan in religious texts stretch over a long period but seem to have been rare throughout. According to the MED s.v. 1b, the earliest examples of collective use of gersum(e) occur in saints’ lives of the twelfth century from Rochester and in the likewise early Lambeth Homilies from the South West of England, from the same region as the Brut. Both collections contain numerous other Norse loans.31 The following example is from the Lambeth Psalter: (16) Đa ileaffullen brohton heore gersum and leiden heo et þere apostlan fotan (Lamb. Hom.Pentec. 101) ‘The believers brought their treasure and laid it at the feet of the apostles’

After these early attestations for gersum(e), the MED s.v. 1b records the noun in a variety of other religious texts until the fifteenth century, the latest example is from Mankind, a morality play from East Anglia dated to ca 1470.32 Alliterative combinations with gold (see MED s.v. 1c) are first attested in Laʒamon’s Brut but are also used in roughly contemporary religious prose texts of the area such as The Wohunge of Ure Lauerd, where gold & gersum is mentioned in a contemplation on the worth of worldly goods.33 The regional distribution of the uses of the loan over a wide non-​southern area seems to agree with the early historical uses in the Peterborough Chronicle dealt with in sections 2 and the later uses in the types of historical fiction dealt with in sections 3.1 to 3.3. But a comprehensive

31 See Treharne (1997: 62, 73, 78) and Dance (2013). The MED records such uses also from Ancrene Wisse and Body and Soul. For the important role of religious texts from the early Middle English period for the survival and development of a vernacular culture after the Norman Conquest see Treharne (2012: 129–136). 32 See Ashley (2010: Introduction). 33 See Thompson (1958: xxii–​xxviii, li–​lix, 20).

Changes of Political Rule and the Changing Use of OE

123

assessment of the variety of uses of the word in religious texts would certainly require a more detailed investigation.

5. Concluding Remarks This study has shown that late OE gærsum(a) ‘treasure’, borrowed from ON gersemi in the historical context of Cnut’s rule, was frequently used in the Anglo-​ Saxon Chronicle until the early twelfth century but then replaced with tresor borrowed from Norman French. ME gersum(e) continued to be used with the same meaning until the fifteenth century but with a clear usage restriction to fictional and religious texts that were in various ways culturally related to the former Danelaw. The observations about this usage restriction with regard to text type seem to be corroborated by the negative evidence of a fourteenth-​century historical text from Lincolnshire, namely from Robert Mannyng’s Chronicle.34 In this text produced for a non-​academic readership from the former Danelaw, Mannyng never used gersum(e) but synonymous tresore a number of times.35

References Allen, Rosamund. 1992. Lawman: Brut. Translated, with an Introduction and Notes. New York: St Martin’s Press. Allen, Rosamund. 1994. “The Implied Audience of Laʒamon’s Brut”. In: Le Saux, Françoise H. M. (ed.): 121–​139. Ashley, Kathleen (ed.). 2010. Mankind. Michigan: Medieval Institute Publications. Baker, Peter S. (ed.). 2000. The Anglo-​Saxon Chronicle: A Collaborative Edition, Vol. 8: MS F. A Semi-​ Diplomatic Edition with Introduction and Indices. Cambridge: Brewer. Barron, W[illiam]. R. J. –​S. C. Weinberg (eds.). 1995. Laʒamon, Brut or Hystoria Brutonum. Edition and Translation with Textual Notes and Commentary. London: Longman. Benson, Larry D. (ed.). 1994. King Arthur’s Death: The Middle English Stanzaic Morte Arthur and Alliterative Morte Arthure. Rev. by Edward E. Foster. (TEAMS: Middle English Texts Series). Kalamazoo: Medieval Institute Publications.

3 4 See Turville-​Petre (1988: 4–​6) and Sullens (1996: 13–​22) for the Latin sources he used. 35 See Sullens (1996: index s.v. tresore).

124

Angelika Lutz

Biddle, Martin –​Birthe Kjølbye-​ Biddle. 2016. “Danish Royal Burials in Winchester: Cnut and his Family”. In: Lavelle, Ryan –​Simon Roffey (eds.): 212–​249. Brook, G. L. –​R. R. Leslie (eds.). 1963, 1978. Laʒamon: Brut. 2 vols. (EETS O.S. 250, 277). London: Oxford University Press. Campbell, Alistair (ed.). 19982. Encomium Emmae Reginae. Cambridge: Cambridge University Press. Clark, Cecily (ed.). 19702. The Peterborough Chronicle, 1070–​ 1154. Oxford: Clarendon Press. Clark Hall, John R. 1969. A Concise Anglo-​Saxon Dictionary. 4th edition, with a Supplement by Herbert D. Meritt. Cambridge: Cambridge University Press. Cubbin, G. P. (ed.). 1996. The Anglo-​Saxon Chronicle: A Collaborative Edition, Vol. 6: MS D. A Semi-​Diplomatic Edition with Introduction and Indices. Cambridge: Brewer. Dance, Richard. 2003. Words Derived from Old Norse in Early Middle English: Studies in the Vocabulary of the South-​West Midland Texts. Tempe, Arizona: Arizona Center for Medieval and Renaissance Studies. Dance, Richard. 2013. “ ‘Tomorʒan hit is awane’: Words Derived from Old Norse in Four Lambeth Homilies”. In: Fisiak, Jacek –​Magdalena Bator (eds.): 77–​127. DOE = Cameron, Angus–​Ashley Amos –​Antonette diPaolo Healey –​Haruko Momma (eds.). 1986–​. Dictionary of Old English, A–​I. Toronto: University of Toronto Press. Douglas, David –​George W. Greenaway (eds.). 19812. English Historical Documents, 1042–​1189. London: Eyre Methuen. Elsweiler, Christine. 2011. Laʒamon’s Brut between Old English Heroic Poetry and Middle English Romance. A Study of the Lexical Fields ‘Hero’, ‘Warrior’ and ‘Knight’. Frankfurt: Lang. Fischer, Andreas. 1989. “Lexical Change in Late Old English: From Æ to Lagu”. In: Andreas Fischer (ed.): 103–​114. Fischer, Andreas (ed.). 1989. The History and the Dialects of English. Festschrift for Eduard Kolb. Heidelberg: Winter. Fisiak, Jacek –​Magdalena Bator (eds.). 2013. Foreign Influences on Medieval English. Frankfurt: Peter Lang. Gameson, Richard (ed.). 2012. The Cambridge History of the Book in Britain, Vol. I: c. 400–​1100. Cambridge: Cambridge University Press. Hanna, Ralph III (ed.). 1974. The Awntyrs off Arthure at the Terne Wathelyn. An Edition Based on Bodleian Library MS. Douce 324. Manchester: Manchester University Press.

Changes of Political Rule and the Changing Use of OE

125

Hanna, Ralph III. 2004. “Alliterative Poetry”. In: David Wallace (ed.): 488–​512. Helsinki Corpus = Rissanen, Matti –​Merja Kytö –​Leena Kahlas-​ Tarkka –​Matti Kilpiö –​Saara Nevanlinna –​Irma Taavitsainen –​Terttu Nevalainen –​Helena Raumolin-​ Brunberg. 1991. The Helsinki Corpus of English Texts. Helsinki: Department of English, University of Helsinki. Hegedűs, Irén (ed.). 2012. English Historical Linguistics 2010. Selected Papers from the Sixteenth International Conference on English Historical Linguistics (ICEHL 16), Pécs, 23–​27 August 2010. Amsterdam: John Benjamins. Herrtage, Sidney J. (ed.). 1879. The English Charlemagne Romances: Sir Ferumbras. EETS, E.S., 97, 99. Horrall, Sarah M. (ed.). 1978. The Southern Version of Cursor Mundi. Vol. I. Ottawa: University of Ottawa Press. Irvine, Susan (ed.). 2004. The Anglo-​Saxon Chronicle: A Collaborative Edition, Vol. 7: MS E. A Semi-​ Diplomatic Edition with Introduction and Indices. Cambridge: Brewer. Keynes, Simon D. 1998. “Supplementary Introduction”. In Campbell, Alistair (ed.): xiii–​lxxxvii. Keynes, Simon D. 2012. “Manuscripts of the Anglo-​Saxon Chronicle”. In: Gameson, Richard (ed.): 537–​552. LALME = McIntosh, Angus –​M. L. Samuels –​Michael Benskin (eds.). 1986. A Linguistic Atlas of Late Medieval English. 4 vols. Aberdeen: Aberdeen University Press. Lavelle, Ryan –​Simon Roffey (eds.). 2016. Danes in Wessex: The Scandinavian Impact on Southern England, c. 800–​c. 1100. Oxford: Oxbow Books. Lawton, David. 1982. “Middle English Alliterative Poetry: An Introduction”. In: Lawton, David (ed.): 1–​19, 125–​129. Lawton, David (ed.). 1982. Alliterative Poetry and its Literary Background. Cambridge: Brewer. Le Saux, Françoise H. M. 1989. Laʒamon’s Brut. The Poem and its Sources. Cambridge: Brewer. Le Saux, Françoise H. M. (ed.). 1994. The Text and Tradition of Laʒamon’s Brut. Cambridge: Brewer. Levy, Bernard S. –​Paul E. Szarmach (eds.). 1981. The Alliterative Tradition in the Fourteenth Century. Kent, Ohio: The Kent State University Press Lucas, Peter –​Angela Lucas (eds.). 2002. Middle English from Tongue to Text: Selected Papers from the Third International Conference on Middle English: Language and Text, held at Dublin, Ireland, 1–​ 4 July 1999. Frankfurt: Peter Lang.

126

Angelika Lutz

Lutz, Angelika. 2012. “Norse Influence on English in the Light of General Contact Linguistics”. In: Hegedűs, Irén (ed.): 15–​41. Lutz, Angelika. 2013. “Language Contact and Prestige”. Anglia 131: 562–​590. Lutz, Angelika. 2018. “König Knuts Schatz”. In: Rohrbach, Lena –​Sebastian Kürschner (eds.): 29–​40. Lutz, Angelika. 2019. “The Uses of Norse Loanwords in Middle English Poems: From Historical Fact to Historical Fiction”. Interdisciplinary Journal for Germanic Linguistics and Semiotic Analysis 24: 21–​60. McNelis III, James I. 1994. “Laʒamon as Auctor”. In: Le Saux, Françoise H. M. (ed.): 253–​272. MED = Kurath, Hans –​Sherman M. Kuhn et al. 1952–​2001. Middle English Dictionary. Ann Arbor, Michigan: University of Michigan Press. Morris, Richard (ed.). 1874–​1892. Cursor Mundi. Four Versions. Parts I–​VI. (EETS O.S. 57, 59, 62, 66, 68, 99). Reprinted 1961–​1966. O’Brien O’Keeffe, Katherine (ed.). 2001. The Anglo-​ Saxon Chronicle: A Collaborative Edition, Vol. 5: MS C. A Semi-​Diplomatic Edition with Introduction and Indices. Cambridge: Brewer. OED = The Oxford English Dictionary. 3rd edition (in progress). Oxford: Oxford University Press. [accessed February 2021] Orel, Vladimir. 2003. A Handbook of Germanic Etymology. Leiden: Brill. Pearsall, Derek. 1981. “The origins of the Alliterative Revival”. In: Levy, Bernard S.–​Paul E. Szarmach (eds.): 1–​24. Pearsall, Derek. 1982. “The Alliterative Revival: Origins and Social Background”. In: Lawton, David (ed.): 34–​53, 132–​136. Peters, Hans. 1981. “Zum skandinavischen Lehngut im Altenglischen”. Sprachwissenschaft 6: 85–​124, 169–​185. Pons-​Sanz, Sara M. 2013. The Lexical Effects of Anglo-​Scandinavian Linguistic Contact on Old English. Turnhout: Brepols. Roberts, Jane. 1994. “A Preliminary Note on British Library, Cotton MS Caligula A.ix”. In: Le Saux, Françoise H M. (ed.): 1–​14. Rohrbach, Lena –​Sebastian Kürschner (eds.). 2018. Deutsch-​isländische Beziehungen: Festschrift für Hubert Seelow zum 70. Geburtstag. Berlin: Nordeuropa-​Institut der Humboldt-​Universität. Skaffari, Janne. 2002. “The Non-​Native Vocabulary of the Peterborough Chronicle”. In: Lucas, Peter–​Angela Lucas (eds.): 235–​246. Skaffari, Janne. 2009. Studies in Early Middle English Loanwords. Norse and French Influences. (Anglicana Turkuensia 26). Turku: University of Turku.

Changes of Political Rule and the Changing Use of OE

127

Sullens, Idelle (ed.). 1996. Robert Mannyng of Brunne: The Chronicle. (Medieval and Renaissance Texts and Studies 153). Binghamton: Binghamton: Center for Medieval and Early Renaissance Studies, State University of New York. Thompson, W. Meredith (ed.). 1958. Þe Wohunge of Ure Lauerd. (EETS O.S. 241). Oxford: Oxford University Press. TOE = Roberts, Jane –​Christian Kay, with Lynne Grundy. 1995. A Thesaurus of Old English. 2 vols. London: King’s College. Townend, Matthew. 2002. Language and History in Viking Age England: Linguistic Relations between Speakers of Old Norse and Old English. Turnhout: Brepols. Treharne, Elaine (ed.). 1997. The Old English Life of St Nicholas with the Old English Life of St Giles. (Leeds Texts and Monographs 15). Leeds: Leeds University Press. Treharne, Elaine. 2012. Living through Conquest: The Politics of Early English, 1020–​1220. (Oxford Textual Perspectives). Oxford: Oxford University Press. Turville-​ Petre, Thorlac. 1988. “Politics and Poetry in the Early Fourteenth Century: The Case of Robert Mannyng of Brunne”. Review of English Studies, N.S. 39: 1–​28. Vennemann, Theo. 2012. “Athel and its Relatives: Origin and Decline of a Noble Family of Words”. English Studies 93: 950–​986. de Vries, Jan. 1977. Altnordisches etymologisches Wörterbuch. Zweite verbesserte Auflage. Leiden: E. J. Brill. Wallace, David (ed.). 2004. The Cambridge History of Medieval Literature. Rev. Paperback Edition. Cambridge: Cambridge University Press. Whitelock, Dorothy (ed.). 19792. English Historical Documents, c. 500–​1042. London: Eyre Methuen.

Rafał Molencki

The Rise of the Verb Happen in Middle English Abstract: The article discusses the field of Medieval English verbs of happening where in Middle English one can observe the gradual lexical replacement of the verbs inherited from Old English (especially (ge)limpan and befeallan) with the new English formations which used the Norse root happ-​. In the late fourteenth and in the fifteenth century happe(n) and happen(en) compete, with the latter prevailing in the end. Both these verbs do not only replace their earlier counterparts in both senses of ‘to have luck to do something’ and ‘to take place’ but also have similar syntactic properties such as the tendency to appear in impersonal constructions (with the original dative subject or Experiencer) and to trigger off subject raising. However, throughout Late Middle English multiple synonymy or lexical layering of verbs of happening can be observed. Keywords: verbs of happening, lexical replacement, Middle English, Old Norse

1. Introduction Early Middle English inherited several Old English verbs that expressed happening of events. The most frequent Old English verb (ge)limpan had become obsolete by the end of the fifteenth century (OED). Verbs of motion such as comen and fallen also served this function, both on their own and in combination with prefixes and/​or phrasal particles and this continued throughout Middle English, e.g. bicomen, comen about/​up/​forth and especially bifallen. All these verbs developed two major senses: ‘to occur, take place’ and ‘to have the fortune, luck to do something’. In Late Middle English all of them came to compete with new formations happen and happenen derived from the noun hap ‘luck, fortune’ borrowed in English from Old Norse in the thirteenth century. As usually happens in such cases, the process of lexical replacement was gradual, which some linguists describe as layering (cf. Hopper 1991, Vanhowe 2008, Allan 2016), defined by Hopper (1991: 22) in the following way: “within a broad functional domain, new layers are continually emerging. As this happens, the older layers are not necessarily discarded, but may remain to coexist with and interact with the newer layers”. Following Traugott (2008), Martín Arista (2011, 2014) and Brems (2012), who use the concept discussing both syntactic and lexical phenomena in the context of grammaticalization, I provide further evidence

130

Rafał Molencki

for the extension of layering to the lexical component given that languages may tolerate multiple synonymies for centuries. The illustrative language material for this study comes from the historical lexical databases: Dictionary of Old English (DOE), Middle English Dictionary (MED), Oxford English Dictionary (OED), Historical thesaurus of the Oxford English dictionary (HTOED) and two corpora: Corpus of Middle English Prose and Verse (CMEPV) and the Shakespearean corpus. The abbreviated titles of the original historical texts follow the conventions of notation used in these sources.

2. Old English Verbs of Happening and Their Middle English Continuations The Historical Thesaurus of the Oxford English Dictionary lists as many as 26 items in the semantic field of verbs of happening, e.g. (ge)limpan, belimpan; feallan, befeallan; (be)cuman; geweorþan; (ge)tidan, (ge)timian; (ge)sælan (the last of which did not even survive into Middle English); cf. also the Old English lexical database project by Martín Arista et al. (2009). As could be expected, most of them are derived from verbs of motion and expressions of time. Some of them are morphologically simple, while others are complex formations made up of two morphemes and there is also one idiomatic expression to mannum/​ manncynne cuman ‘come to men/​ mankind’. Using the Natural Semantics Metalanguage Model Mateo Mendaza (2016) proves that HAPPEN is a semantic prime based on her study of Old English verbs of happening. The Old English verbs of happening mostly survived in Early Middle English, but many of them had become obsolete by the end of the Middle Ages. The sections below provide some examples.

2.1. OE (Ge-​)limpan, Belimpan > ME Limpen, Bilimpen This was a word inherited from Proto-​Germanic *lempan (ultimately from the reconstructed Proto-​Indo-​European root (s)lembʰ-​ ‘to hang limply’), as confirmed, by cognate early Germanic forms, e.g. Old High German limpfan, Old Saxon limpan, Old Frisian limpa (cf. Pokorny 1959: 655). It was a common Old and Early Middle English verb: (1) (2)

He hiene […] ofslog, swa him eac selfum siþþan æfter lamp. (Or 4 5.91.23) ‘He killed him, as also happened to himself afterwards’. Hit on endestæf eft gelimpeþ ðæt se lichoma læne gedreoseþ. (Beowulf 1754) ‘Then finally the end arrives [literally: it happens that] when the body he was lent collapses’. (Heaney 2000: 120)

The Rise of the Verb Happen in Middle English (3)

(4)

(5)

131

Nu we willen sægen sumdel wat belamp on Stephnes kinges time. (ChronE 1137.71) ‘Now we will say something about what happened in king Stephen’s time’. Ȝif hit [a child] is misborn, as hit ilome limpeð [Bod: ilimpeð], & wont eni of his limen oþer sum misfare, hit is sorhe to hire. (a1250 HMaid.(Tit D.18) 31/​503) ‘If a child is born disabled, as often happens, and one of its limbs is missing or has some kind of defect, it is a grief to her [its mother]’. And vneth limpyd hym þe lyfe. (c1450(?a1400) Wars Alex. 2060) ‘And an uneasy life happened to him’.

According to the OED data, the last instances of limpen are attested in the early fifteenth century while belimpen is only found in the Early Middle English texts that have their sources in Old English.

2.2. OE Feallan, Befeallan > Middle English Fallen, Bifallen Both fallen and befallen were found as synonyms of (be)limpen in Medieval English and the latter continued to be used in Early Modern English and even later in archaized texts. (6)

(7)

(8) (9)

se hælend forðon uiste alle ða ðe toweardo uoeron ofer hine feoll ł cuom ł forðeode & cuoeð him huoelcne soecas gie (Ru gifeoll, CpH eode [...] forð). (JnGl (Li) 18.4) ‘The saviour therefore, knowing all things that should come upon him, went forth, and said unto them, Whom seek ye?’ ðus he mid tearum biddende, him eft oþer geþanc on befeoll þus cweðende. (LS 23 (Mary of Egypt) 794) ‘As he was praying thus with tears again another thought happened [came into his mind], saying thus’. Hit fel in one daye, þe king was out iwent. (?a1300 Jacob & J.(Bod 652) 193) ‘One day it happened that the king went out.’ Bifil that in that seson on a day In Southwerk at the Tabard as I lay […] At nyght were come into that hostelrye Wel nyne and twenty in a compaignye. (c1387–​95 Chaucer CT.Prol. A.19) ‘It happened that in that season on one day, In Southwark at the Tabard Inn as I lay…At night had come into that hostelry Well nine and twenty in a company.’1

2.3. OE (Be)cuman > ME (Bi)comen Bicomen when used as a verb of happening often collocated with inevitable events such as death and God’s judgment:

1 Capitalization as in the Riverside Chaucer Benson’s (2008) edition, preserved in the MED.

132 (10)

(11)

(12)

Rafał Molencki […] us becom ða (eað. & forwyrd. þurh wif: & us becom eft lif. & hredding þurh wimman. (ÆCHom I,13 283.72) ‘Death and perdition befell us through a woman, and afterwards life and salvation came to us through a woman’. […] ða bicom heom feringa […] þæt heoræ naðor nan word cwæðen ne mihte. (LS 5 (InventCrossNap) 120) ‘Then it suddenly occurred to them that neither of them could say a word’. Seoððen his ende bicom. (c1275(?a1200) Lay. Brut (Clg A.9) 10850) ‘Later his end happened’.

In later Middle English bicomen began to narrow down its sense to the modern sense of the verb of becoming replacing the obsolescent (i)worthen in this role. The high-​frequency verb of motion comen also developed the sense equivalent to ‘to happen’: (13) (14) (15)

[…] we nyton hwylce dæge oþðe on hwylce tid se deaþ cymeð. (HomM 11 62) ‘We do not know on which day or at what time the death will come/​happen.’ Þa com an mycel storm & to dræfede ealle þa scipe. (a1121 Peterb.Chron.(LdMisc 636) an.1070) ‘There arose/​happened a great storm and dispersed all the ships.’ Al that comth, comth by necessitee. (a1425(c1385) Chaucer TC (Benson-​ Robinson) 4.958) ‘Everything that happens, happens of necessity.’

In his manual Byrhtferth used idiomatic expressions cuman to mannum/​ mancynne in the sense of ‘occur, happen (to people)’: (16)

[…] and autumnus, þæt ys hærfesttima, cymð to mancynne binnan seofon nihta fyrste. (ByrM 1 2.1.332) ‘and autumn, that is harvest-​season, comes/​happens to mankind after a period of seven days.’

In Late Middle English comen in the sense of ‘to happen’ came to be used with prepositional particles about, forth, (up)on: (17)

He meruelled hou it cam aboute. (a1450 Gener.(1) (Mrg M 876) 8775) ‘He wondered how it came about/​happened.’

2.4. OE (Ge)weorþan > ME (I)worthen Late Middle English (i)worthen became obsolete both in the sense of becoming (being replaced with bicomen) and happening. All the instances of the verb meaning ‘to happen’ cited in the MED date from Early Middle English only: (18)

eac on þæm geare gewurdon monega wundor on monegum londum. (Or 10.123.13: 7) ‘And also in this year many wonders happened in many lands.’

The Rise of the Verb Happen in Middle English (19)

(20)

133

Ðises geares gewurdon swiðe fela þunra. (a1121 Peterb.Chron. (LdMisc 636) an.1109) ‘There happened much thunder this year.’ Al hit scal iwurðe [Otho: iworþe] þus. (c1275(?a1200) Lay. Brut (Clg A.9) 13964) ‘All of it will happen thus.’

2.5. OE (Ge/​Be)tidan, (Ge/​Be)timian > ME (I/​Bi)tiden, (I/​Bi)timen Another group of verbs that denoted happening was derived from the nouns referring to time: tide ( 24/​1,000 wds). However, he is also keen to provide his thoughts on the texts he is using and on the text he is producing, in the form of assessment markers (f > 28/​1,000 wds), which are often reinforcing, not down-​ toning. This position is corroborated by use of modal verbs that either indicate obligation or inevitability: the author uses absolute values rather than moderate values in his propositions. This is at times counterbalanced by hedges, but they are very few, and thus their effect throughout the text minimal. Finally, code glosses position the author somewhat above the expected audience, and as they do not occur in the Latin original (see the discussion in Harvey –​Tavormina 2020: 285–​288), it can be assumed that the intended audience of the ME translation of the prologue is somewhat different from that of the Latin version.

Figure 1:  Metadiscourse items in LU prologue, normalised frequencies

Rhetorical Re-analysis of Metadiscourse Items

183

6. Reanalysis of Metadiscourse Items This section will take a look at three metadiscourse classes, i.e. self-​references, evidentials, and attitude markers. The selection is limited to these classes as firstly, they offer several items each for the analysis, and secondly, the aim of the paper is to introduce and exemplify the method, and therefore it is purposeful not to make the reanalysis overly complex. The aim is to reanalyse the items in the three classes in the rhetorical plane defined by the concepts ethos, pathos and logos. The reanalysis of individual metadiscourse items is carried out with respect to the cotext and function of the linguistic items, paying attention to the situation in which the items are used. This requires the scrutiny of the passage in which an item occurs. Metadiscourse items are then allocated to the three modes of persuasion in each case. Eventually we can map the occurrences of metadiscourse items onto the plane of rhetorics. It is worth noting here that in this analysis, there can be passages that do not belong clearly to only one type of persuasion. Therefore, we are looking at plane of persuasion where a passage may have features of one or more of the rhetorical concepts, and thus we are no longer allocating the metadiscourse items to nominal rhetorical variables; rather, we are transforming one metadiscourse item class into a scalar variable on a continuum joining ethos, pathos and logos.

6.1. Mapping Self-​references onto a Rhetorical Plane There are six distinct types of self-​references in the LU prologue. The following list of the types also provide the analysis of the types into their kinds of persuasion. 1. Plain identification of the author (ethos) 2. Explanation of method/​choices made (ethos/​logos) 3. Difficulties or expressions of obligation (ethos/​pathos) 4. Description of challenging situations (ethos) 5. Claim-​making, opinionated text (ethos/​logos) 6. Expression of wish or request (ethos/​pathos). The class of self-​reference has 32 occurrences in the prologue, most often in the form of word I. The cotext they occur in gives us clues as to how one should analyse each occurrence. All the self-​mentions rely more or less on ethos. Items in 1) set the manner in which the author wishes to be identified, i.e. the persuasiveness relies on ethos, e.g. Frere Henry Danyel of the ordre of ffrereȝ prechoures þe seruant of Jhesu crist & of þe virgine his moder. Items in 2) are based on ethos and logos, i.e. the passages appeal often to the reason of the audience (even if the author may also be presented in negative light as well); also, many of the passages

184

Martti Mäkinen

are related to the explanation of choices made in the translation work, and in that references to texts and authors should, naturally, be allocated to logos (ffor which causes more þan for scarsenes of þe langage I lefte meny þingȝ or addede to . myne owen þat som auctoures affermeth noȝt). Items in 3) rely in addition to ethos also on pathos, as the difficulties and obligations described may wake the readers’ feelings of compassion (And þat I shulde write ye it shortly). 4) makes good use of ethos: we know the author has succeeded in the translation and compilation work, after all, we read his prologue to the work (And also for y haue noȝt mynde that I haue redde ne harde neyþer yis science giffen in English). In 5) we see claims or opinions expressed, and in that the author assumes a position in which he is entitled to utter his thoughts, thus in addition to the appeal to reason, also ethos is strongly present (And forþi þat euery þing in how mich it is more openly taghte in so myche it schal be take more liȝtly & of moo men þerfor þat I be noȝt made a liȝtede lanterne to hide vnder a busshel but þat I condescende vnto þi praiers & noȝt only to þine & þam þat be like þe but þat I eke & encresce in gode þe knowlich of al þam þat coueteþ forto profite in þis faculte). Items in 6) rely on pathos (at least in the eyes of the modern audience), and thus, despite the clear authorial presence, this passage can also be deemed emotional (Therfor in þis werk I aske euermore a trewe reder I desir a meke herer I biseke euermore a lele & a benigne correctour). The cotexts of the occurrences of self-​references will help us in assessing how “Daniel” actually uses this metadiscourse class: the focus is not only on ethos, even if that is almost always unavoidable. Thus, if the self-​reference points are mapped onto the plane of rhetorics (Figure 2), we can see how they do not amass around the factor “ethos”, but are more evenly distributed over the field, touching also upon “pathos” and “logos”. The metadiscourse items for evidentials and attitude markers are mapped in Figure 2 in a similar manner. The lists below provide the different metadiscourse item types, their numbers (also used in Figure 2), and how they have been analysed with respect to the rhetorical concepts.

Evidentials 7. Authorities by name (logos) 8. References to Bible (logos/​pathos) 9. Authorities generally (logos/​pathos) 10. Daniel/​self (ethos/​logos) 11. Textual references by title (logos)

Rhetorical Re-analysis of Metadiscourse Items

185

Attitude markers 12. Author in difficult situation (ethos/​pathos) 13. Statement of fact/​claim (logos) 14. Opinion, confirmation, assertive statement (ethos/​pathos) 15. Description of request (pathos)

Figure 2:  Metadiscourse items reanalysed on rhetorical plane Legend: ∆ = self-​reference item ◊ = evidential item ▲ = attitude marker. Figures in triangles: the metadiscourse item type as indicated in text.

In Figure 2 (NB: a two-​ dimensional plane, not a projection of a three-​ dimensional space), we can see the 15 different metadiscourse types (numbered according to the lists in this section) from three different metadiscourse categories (indicated by different shapes) mapped onto rhetorical plane defined by ethos, pathos, and logos. Each instance of the different types has been positioned on the plane, and their respective positions reflect the outcome of the reanalysis of the metadiscourse items. If an item can be understood to represent more than one type of persuasion, then it has been placed in the middle of the relevant types, like type 8, references to Bible, which can be understood to appeal to both reason and emotions, thus the position between pathos and logos. This means that the analysis does not consider the intensity of the item’s persuasive force

186

Martti Mäkinen

within one or the other rhetorical concept, and thus the analysis positions an item between the two poles. The circles in Figure 2 represent the average of the metadiscourse class on the persuasive plane: for example, there are almost as many “ethos” items among self-​references as there are “ethos/​pathos” items, or “ethos/​logos” items. Therefore, the average of all self-​references is placed near the “gravitational centre” of all self-​reference items on the plane. The averages for evidentials and attitude markers have been placed in a similar manner. Finally, the core of “Daniel’s” persuasive performance has been demarcated by strong, dashed lines, enfolding the three averages. Looking at this core one can infer that even if the author uses all the walks of Aristotelian persuasion in the LU prologue, he is leaning slightly more towards ethos and logos than to pathos. At the same time, one must bear in mind that this result has been achieved through the analysis of a selection of available items that can be, to some extent, pinned on linguistic categories (self-​references, evidentials, and attitude markers), and therefore it does not provide a full model of all the persuasive means used in the prologue. If the analysis had regarded e.g. modal verbs, the view of the persuasiveness and the persuasive means of the prologue would have been different. Nevertheless, the analysis provides an example of extending the rhetorical analysis into linguistic items, despite the restrictions set by the sample used.

7. Discussion The two goals of this article, describing the persuasive means of the LU prologue in terms of metadiscourse, and attempting a move towards a qualitative-​ quantitative linguistic analysis of persuasion have been presented above. The analysis shows the inventory of the “Daniel’s” metadiscourse items, and it also provides a re-​analysis of a selection of them on a rhetorical plane. The selection obviously restricts our view of the author’s use of the different facets of classical rhetorics, but the purpose of the re-​analysis is more to show the usability of the method, not to provide an exhaustive picture of “Daniel’s” rhetorical competence. Both the metadiscourse and rhetorical analyses are based on qualitative, intuitive close-​reading of the text. Nevertheless, in the analysis of metadiscourse items, some linguistic categories are identified, and that provides an inventory of items that can be transferred over to the rhetorical analysis, thus providing a linguistic base for the investigation of persuasion. Also, looking at the metadiscourse item inventory through the lense of rhetorics and the different classical ways of appealing to one’s audience gives us an idea of the layered characteristics and

Rhetorical Re-analysis of Metadiscourse Items

187

the multiple functions of the items identified in the earlier analysis: many of the metadiscourse items looked at in this article belong to more than one classical concept of persuasion. As such, this method may provide insights into metadiscourse analysis for scholars unfamiliar with it, and vice versa into the rhetorical concepts of persuasion. The stacked nature of the analysis also requires some thought: the method begins with a qualitative analysis of metadiscourse items, the results of which are fed into the qualitative analysis of rhetorical concepts, which, finally, is rendered into a quantitative analysis of the author’s persuasive performance. It is justified to question the reliability and reproducibility of such an analysis and apparatus. As with any investigation, also here the key is the documentation of the consecutive steps of the study: only that will lessen the effect of the subjective elements of the study. For this end, also the transcription of the prologue is provided in Appendix 1. Hyland’s metadiscourse inventory excludes many other persuasive features in texts, one of the obvious ones being the choice of words and the nuances an author attempts to convey with them. Often the metadiscursive items do not “enter the syntax”, i.e. leaving them out would in many cases still leave us a syntactically sound sentence. Therefore, many of the metadiscourse items, e.g. boosters and transition/​frame markers are something extra in a sentence, and therefore Hyland’s model does not reach the level of persuasion that takes place in the essential sentence constituents. Also, Hyland does not address the question whether all persuasive elements are equally persuasive. Here Östman’s notion of more or less persuasive texts can be scaled down to item level, and we can conclude that no, not all items are equally persuasive. This conclusion obviously has repercussions on the quantitative re-​analysis of metadiscourse items, but it is not in the scope of this article. In the same vein, the study of historical texts does not allow us access to the contemporary audience reactions (the exception, of course, is correspondence in cases when the letters of both parties have survived). This entails that we cannot quantify the persuasive effect from the point-​of-​view of the audience. Nevertheless, we can allow ourselves an informed guess at the authorial persuasive intention, provided that we are familiar with the community and era that produced the text. Therefore, any analysis of an author’s persuasive performance must be performed against our knowledge of the historical context. That entails that the analysis is as reliable as our knowledge of the culture and the persons who produced the text. Finally, there remains the identity of the translator of the LU prologue into Middle English: Harvey –​Tavormina (2020: 285, 307–​308) put forward the idea

188

Martti Mäkinen

that the translator of the prologue is someone other than Henry Daniel, based on stylistic differences and some differences detected between the dialectal variants of ME in the prologue and in the main text. This is corroborated by the use of code glosses in the ME version: Walter Turner of Ketton, to whom the original Latin prologue was addressed, did not need the ME translation to understand Daniel’s prologue, let alone the glosses to explain the Latinate terms used (­examples 20–​22). Nor did Daniel have a motive to translate the prologue into English, as the prologue was not intended to the same audience as the main text of LU. Therefore, it seems safe to assume that the translator of the ME prologue is not Henry Daniel.

8. Conclusions This article set out to describe the persuasive means used in the Middle English prologue to Henry Daniel’s Liber uricrisarum, and to provide an example of using linguistic items in a rhetorical analysis of the text. According to the metadiscourse items re-​analysed, Daniel’s prologue seeks to engage the readers, and to reason with them, preferring ethos and logos to pathos. As Daniel was among the first to publish a text of this magnitude and sophistication in English, he must have felt the pressure to justify it. The study shows the usability of Aristotelian rhetorical in the analysis of an inventory of metadiscourse items, and in that it can provide a way to triangulate the persuasive means used in a text. The proposed method is a step forward; nevertheless, it requires development. As the focus here is on a translated, historical text, this particular study will benefit from the support of translation studies, and especially of an analysis of the paratextual functions of the prologue. In this case, at least a closer comparison of the Latin original and ME translation should be carried out, in order to be able to differentiate Daniel’s persuasive means from the translator’s: already in the course of this study we have seen that not everything was carried over from Latin to English, and some elements were added.

Rhetorical Re-analysis of Metadiscourse Items

189

References Alonso-​Almeida, Francisco –​Mercedes Cabrera-​Abreu. 2002. “The Formulation of Promise in Medieval English Medical Recipes: A Relevance-​Theoretic Approach”. Neophilologus 86: 137–​154. Chaemsaithong, Krisda. 2011. “Discursive control and persuasion in early modern news discourse: The case of English witchcraft pamphlets and their prefaces (1579–​1621)”. English Text Construction 4:2: 228–​256. Corbellini, Sabrina –​Margriet Hoogvliet –​Bart Ramakers (eds.). 2015. Discovering the Riches of the Word. Leiden: Brill. Dearnley, Elizabeth. 2016. Translators and Their Prologues in Medieval England. (Bristol Studies in Medieval Cultures). Cambridge: D.S. Brewer. Getz, Faye Marie. 1990. “Charity, translation and the language of medical learning in medieval England”. Bulletin of the History of Medicine 64/​1: 1–​17. Gotti, Maurizio –​Marina Dossena (eds.). 2001. Modality in Specialized Texts. Bern: Peter Lang Halmari, Helena –​Tuija Virtanen. 2005. “Persuasion across genres: Emerging perspectives”. In: Halmari, Helena –​Tuija Virtanen (eds.): 3–​24. Halmari, Helena –​Tuija Virtanen (eds.). 2005. Persuasion Across Genres: A Linguistic Approach. Philadelphia: John Benjamins. Harvey, E. Ruth –​Teresa Tavormina. 2020. Liber Uricrisiarum: A Reading Edition. Toronto: University of Toronto Press. Hanna III, Ralph. 1991. “Chaucer’s Meter: The Evidence of the Manuscripts”. In: Machan, Tim William (ed.): 17–​40. Hanna III, Ralph. 1994. “Henry Daniel’s Liber Uricrisiarum (Excerpt)”. In: Matheson, Lister (ed.): 185–​218. Hyland, Ken. 2009. Academic Discourse: English in a Global Context. London: Continuum International. Hyland, Ken. 2005. Metadiscourse: Exploring Interaction in Writing. London and New York: Continuum. Hyland, Ken. 1998. “Persuasion and context: The pragmatics of academic metadiscourse”. Journal of Pragmatics 30, 437–​455. Jasin, Joanne. 1983. “A critical edition of the Middle English Liber uricrisiarum in Wellcome MS 2–​25”. Diss. Tulane University. Jasin, Joanne. 1993. “The transmission of learned medical literature in the Middle English Liber uricrisiarum”. Medical History 37/​3: 313–​29.

190

Martti Mäkinen

Jucker, Andreas H. 1997. “Persuasion by inference: Analysis of a party political broadcast”. In: Blommaert, Jan –​Chris Bulcaen (eds.). Political Linguistics. Belgian Journal of Linguistics 11: 121–​137. Kinneavy, James L. 1971. A Theory of Discourse: The Aims of Discourse. Englewood Cliffs, NJ: Prentice-​Hall. Kirk, John M. (ed.). Corpora Galore: Analyses and Techniques in Describing English. Amsterdam: Rodopi. Levere, Trevor H. (ed.). 1982. Editing Texts in the History of Science and Medicine: Papers Given at the Seventeenth Annual Conference on Editorial Problems University of Toronto, 6–​7 November 1981. New York: Garland Publishing. Liira, Aino –​Sirkku Ruokkeinen. 2019. “Material Paratext Studies: Redefining the Concept of Text in Light of Manuscript Evidence”. Anglicana Turkuensia 33: 111–​134. https://​www.utupub.fi/​handle/​10024/​148123 [accessed May 2020]. Machan, Tim William (ed.). 1991. Medieval literature: Texts and Interpretation. (Medieval and Renaissance Texts and Studies 79). Binghamton: Center for Medieval and Early Renaissance Studies. Mäkinen, Martti. 2011. “Swaying the medical audience: persuasion in early English medical instructional passages”. Poster presented at ICAME 32, June 1–​6, 2011, University of Oslo. Marttila, Ville. 2011. “New arguments for new audiences: a corpus-​ based analysis of interpersonal strategies in early Modern English medical recipes”. In: Taavitsainen, Irma –​Päivi Pahta (eds.): 135–​157. Matheson, Lister (ed.). 1994. Popular and Practical Science of Medieval England. East Lansing: Colleagues Press. Nevalainen, Terttu –​Leena Kahlas-​ Tarkka (eds.). 1997. To Explain the Present: Studies in the Changing English Language in Honour of Matti Rissanen. Helsinki: Société Néophilologique. Nwogu, K.N. 1997. “The medical research paper: Stucture and functions”. English for Specific Purposes 16(2): 119–​138. Östman, Jan-​ Ola. 2005. “Persuasion as implicit anchoring: The case of collocations”. In: Halmari Helena –​Tuija Virtanen (eds.): 185–​212. Pahta, Päivi 1998. Medieval Embryology in the Vernacular: The Case of De Spermate. Diss. Helsinki: Société Néophilologique. Pahta, Päivi –​Irma Taavitsainen. 2004. “Vernacularisation of Scientific and Medical Writing in its Sociohistorical Context”. In: Taavitsainen, Irma –​Päivi Pahta (eds.): 1–​19.

Rhetorical Re-analysis of Metadiscourse Items

191

Peikola, Matti. 2018. “ ‘Hit is false poynted in many places’: Paratextual communication on punctuation in early English prologues, title-​pages and errata lists”. In: Kytö, Merja –​Bo Andersson (eds.). Punctuation: Past and Present. Special issue of Studia neophilologica 90 supplement 1: 50–​67. Peikola, Matti. 2015. “Manuscript Paratexts in the Making: British Library MS Harley 6333 as a Liturgical Compilation”. In: Corbellini, Sabrina –​Margriet Hoogvliet –​Bart Ramakers (eds.): 44–​67. Peikola, Matti –​Janne Skaffari –​Sanna-​Kaisa Tanskanen (eds.). 2009. Instructional Writing in English: Studies in Honour of Risto Hiltunen. Amsterdam, NLD: John Benjamins. Quintana-​ Toledo, Elena. 2009. “Middle English Medical Recipes: A Metadiscursive Approach”. Studia Anglica Posnaniensia 45/​2: 21–​38. Siraisi Nancy G. 1990. Medieval and Early Renaissance Medicine: An Introduction to Knowledge and Practice. Chicago: The University of Chicago Press. Slack, Paul. 1979. “Mirrors of Health and Treasures of Poor Men: The Uses of the Vernacular Medical Literature of Tudor England”. In: Webster, Charles (ed.): 237–​273. Swales, John M. 1990. Genre Analysis: English in Academic and Research Settings. Cambridge: Cambridge University Press. Swales, John M. 2004. Research genres: Explorations and applications. Cambridge: Cambridge University Press. Taavitsainen, Irma 2004. Genres of Secular Instruction: a linguistic history of useful entertainment. Miscelánea: A Journal of English and American Studies 29: 75–​94. Taavitsainen, Irma, 2001a. Middle English recipes: Genre characteristics, text type features and underlying traditions of writing. Journal of Historical Pragmatics 2.1: 85–​113. Taavitsainen, Irma 2001b. “Evidentiality and scientific though-​styles: English medical writing in Late Middle English and Early Modern English”. In: Gotti, Maurizio –​Marina Dossena (eds.): 2–​52. Taavitsainen, Irma. 2000. “Metadiscursive Practices and the Evolution of Early English Medical Writing 1375–​1550”. In: Kirk, John M. (ed.): 191–​207. Taavitsainen, Irma –​Pahta, Päivi 1997. “The Corpus of Early English Medical Writing: Linguistic variation and prescriptive collocation in scholastic style”. In: Nevalainen, Terttu –​Leena Kahlas-​Tarkka (eds.): 209–​25. Taavitsainen, Irma –​Päivi Pahta (eds.). 2004. Medical and Scientific Writing in Late Medieval English. Cambridge: Cambridge University Press.

192

Martti Mäkinen

Taavitsainen, Irma –​Päivi Pahta (eds.). 2011. Medical Writing in Early Modern English. Cambridge University Press. Talbot, Charles H. 1967. Medicine in Medieval England. London: Oldbourne Book Company. Virtanen, Tuija –​Helena Halmari 2005. “Persuasion across genres: Emerging perspectives”. In: Halmari –​Virtanen (eds.): 3–​26. Voigts, Linda Ehrsam. 1982. “Editing Middle English Texts: Needs and Issues”. In: Levere, Trevor H. (ed.): 39–​68. Webster, Charles (ed.). 1979. Health, medicine, and mortality in the sixteenth century. Cambridge: Cambridge University Press.

Appendix 1: Middle English Prologue to Henry Daniel’s Liber uricrisiarum, BL Royal MS 17.D.1, ff. 4r–​5r This prologue was transcribed in the British Library, and the transcription has been checked against that provided in Harvey –​Tavormina (2020: 35–​37). This transcription provides a diplomatic rendition of the original text (as opposed to the reading version of the afore-​mentioned edition), maintaining the manuscript word breaks and differentiation of capital and lower case letter, and also indicating abbreviations and superscript letters in italics. Any remaining mistakes are mine. Þis prolog was first made in latyne . bot after ward it was translate in to ynglisch –​Frere Henry Danyel of the ordre of ffrereȝ prechoures þe seruant of Jhesu crist & of þe virgine his moder . Vnto his belouede felowe in crist Walter Turnour of Ketoun . Moste belouede felowe þou haste praiede me oftymeȝ & besily . þat y shulde gadre to þe one handeful of floureȝ . Of þe domes of vrines . And þat I shulde write ye it shortly . and that in wlgare i. comune langage . Þe which forsoþe to be done . þou haste put me to an harde wark . and opne to the barking & to the cornyng of detractours . Bothe for that þat is gyffen into techis by wrytyng . but it be softening þe eres of men and passing þe witte of many men . it is al arectede into scorn And also for y haue noȝt mynde that I haue redde ne harde neyþer yis science giffen in English , And also for y am nouþer witty ne wise of this tonge . And also for after Aueroys and it is had of Gilbert in his coment vpon Giles that this faculte mai noȝt be schewede by tonge . And also for after ffor whie as it is opne schewede to konnyng men no science may sufficiently be schewede in þis tonge s. english . And þis as y trowe is þe resoun . ouþer for the langage is vnsufficiant in it selfe or for þat we kane it not perfitly . And also for noþer haue I take it ne lerede it of man but as in oþer sciences of interpretacions by þe gifte of the holy gost which departeȝ to euery man . Vnto helpe as he will . But certeynly riȝt wel belouede broþer loue & charite ouercommeþ simple men

Rhetorical Re-analysis of Metadiscourse Items

193

. but couetouse of temporal lucre or of praysyng or of fauour ouercometh hem þat bene childer of þis worlde . And what so eny man coueyteþ moste þat praiseþ he moste ¶ We rede in scriptures meny holy men for þair saiingeȝ & þair writyng ful greuousely haue bene punysshede & suffrede meny eueles . but noȝt but of euel men . Also holy writte makeþ myende þat þe selfe aposteles of crist were arectede reprouabli . Wiþ outen letres and ydiotes ¶ þerfor if so grete men & so holy men ȝa þat were inspirede wiþ þe holy goste myȝt no waies eschewen the contumeleȝ & myssayingȝ or vpbraidinges of proude men scornyngȝ of presumptuouse men . and þe setting at noȝt of enuyouse men & corrupt in soule . And þe hissingȝ & det detracciouns of serpentes tonges . how shal I þat am so litle as þe leste of þe seruantȝ of criste . and þe first article as I trowe þat techeþ þis faculte in english tonge . how shal I inow eschewe euel spekingȝ & vpbraidyngȝ certeynly in no waieȝ ffor I do it noȝt for cause of lucre of favour as oþer men doþ þat þai be sene wise onely in myche speche of speking men þat bene nowe . But certeynly while þat þe euyles . þat ar done to gode men of euel men ar gadrade into þe merite to þe same gode men . gode warkes ar demede of discrete men noȝt to be wiþ drawen for þe maneres of schrewede men ¶ Therfor dere frende I considerand meny men & diuerse þat couaiteþ to be experte of demyng of vrines –​fforþi þat þe science is faire & wonderful and also sich a science is as it were properly myche profitable vnto men And also I seand ful meny men ȝe as it were alle men languishand doutously aboute þe sothfastnesse or trewþe of it . And forþi þat euery þing in how mich it is more openly taghte in so myche it schal be take more liȝtly & of moo men þerfor þat I be noȝt made a liȝtede lantern to hide vnder a busshel but þat I condescende vnto þi praiers & noȝt only to þine & þam þat be like þe but þat I eke & encresce in gode þe knowlich of al þam þat coueteþ forto profite in þis faculte I haue gadrede as I myȝt wiþ grete labour by . 3 . ȝere þis presente werk of þe bokes of meny auctoures & þe sayingȝ of þe comentours of þam . ofte tymes lettede or take þerfro noȝt only for þe labour of þe obediens of myn ordre but also for diuerse infirmites and somtyme almost to þe deth . ffor which causes more þan for scarsenes of þe langage I lefte meny þingȝ or addede to . myne owen þat som auctoures affermeth noȝt ¶ In whiche werke forsoþe ar schewede certeynly rewles of demyng of vrynes after giffyng & teching of auctours and of comentoures of þis sciens wiþ diffiniciouns and exposiciouns of termes of sekenes or infirmites & of membres wiþout forth & wiþinforþe & wiþ meny oþer þingeȝ notable in þis crafte which I demede to be callede Liber Vricrisiarum þe boke of demyng of vryn ¶ fforsoþe he þat biholdeþ wel and perfitely þis boke made in wlgare i. in comune tonge he shal mowe be a parfite domesman in this crafte . And wiþ oute doute he schal gete gode & richeȝ and helthe of his soule . neuerþeles if he mysvse a pompouse lif or bostful . ful of wordes ful of fables & ful of lesingeȝ as leches þat bene now . ar wonte to done ¶ Wise Aristotel forsoþe as it is openly schewede ex fine primi & principio secundi Methaphisice he þankeþ þaym þat bifore him gaf or taght in writyngȝ noȝt only wisely .

194

Martti Mäkinen

but also þam þat dede noȝt wisely for he fonde meny more subtil þingȝ of þo þinges þat oþer wrote bifore ffor after þer same Aristotel in fine 2i libri elenchorvm þo þinges þat ar fonden labored afore of oþer men ar particulerly ekede or encresed of þam þat toke þam afterwarde & sequitur þo þingȝ forsoþ þat ar first fonden ar wont forto take but litil encresyng in þe first tymes but þe encresyng þat is done afterward of þam is more profitable hec ille ffor whie I saie noȝt þis of myself þat I am to write newe þinges but þat I do it newely . I haue gadrede now late for þam þat kan take it in latyn a schorte tretice conteynyng fully þe marowe of þis faculte and I wrote it in þe tong þat forsoþe is riȝt dere to me ¶ Ne wonder noȝt eny man it I sewe noȝt medicynes perteynyng to þe sekenes & þe infirmites tochede in this boke . ffor god helping and þe obedience of myn ordre noȝt agayn standing it bihete me forto make a werk by it selfe þerto ¶ Therfor in þis werk I aske euermore a trewe reder I desir a meke herer I biseke euermore a lele & a benigne correctour ffor charite broght me more herto þan hardines Ouer þat I praie euery writer or compiler of þis þat he kepe my writing . but if he be of þe langage of anoþer contre ffor whie as for þe langage of englissh tong as anentȝ a discrete man & him þat hath þe gift of tunge . trewe & parfite craft of ortographie is taugh in þis bok he þat vnderstondes it noȝt praie he þat he maye interpretate it seiþ þe trompe of criste i. seynt poule ¶ therfor þis present tretis of englisshe is diuidede into 3 . bokes or elle . 3 . particules In þe j . bok is taught of vryn wherof it is saide & what it is wherof and how it is gendrede what þinges and how meny ar to be considerede of a leche & howe þat he owe to haue him self in demyng & it hath . 4 . capitula ¶ In þe 2 . bok of domes & significacions of coloures & of bodies of vryn & it hath . 14 . capitula ¶ In þe . 3 . boke forsoþ of domes & significacions of contenteȝ of vrynes & it hath . 20 . capitula In þe laste of which . 20 . capitula it is tretede of þe rewles of ysaac which him self giffeþ in þe . 10 . partie of his boke of vryns Explicit prologus

Gjertrud F. Stenbrenden

The Evidentiary Status of Back Spellings and English Historical Phonology Abstract: This paper examines back spellings (analogical spellings) in the history of English in order to (a) identify their rationale and (b) determine their evidentiary status with respect to sound-​change. Traditionally, back spellings have always been dismissed as evidence of secondary sound-​change because they have been believed to result from phonemic merger in all cases. However, analysis of a number of historical back spellings, applying the concept of “literal substitution sets”, reveals that the motivation for such spellings is manifold but does not usually involve phonemic merger. Reasons include analogy, calligraphic changes and orthographic extensions, phonetic and phonological innovations, and also phonemic merger. Determining the evidentiary status of back spellings additionally requires detailed assessment of scribal behaviour and of any clashes between scribal and authorial systems, both written and spoken. Keywords: back spellings, Middle English phonology, literal substitution sets, sound-​ change, written vs spoken language

1. Introduction The aim of the present paper is to examine so-​called “back” or “reverse/​inverse/​ analogical” spellings: to assess historical examples of back spellings, mostly from the history of English; to identify the varying rationale for back spellings; and, crucially, to determine when back spellings may be used as evidence for sound-​ change. No empirical evidence has been specifically harvested for this paper, as it seeks to throw light on a long-​standing methodological concern rather than give an empirical contribution to the field of historical English phonology. To elucidate the nature of back spellings, two examples may suffice at this point. Penzl (1957 [1969: 18]) adduces an example from High German, in which the front rounded vowels /​ö/​and /​ü/​unrounded in the twelfth century, causing reverse spellings for the pre-​existing front unrounded vowel phonemes /​i/​and /​e/​. The stages of the process, and the logical motivation behind the subsequent irregular spellings, may be laid out as follows.

196 (1)

Gjertrud F. Stenbrenden 1. HG /​ö/​[ø] > /​e/​ and /​ü/​[ʉ] > /​i/​    = Phonetic unrounding and phonemic merger 2. for historical /​ö/​, and for historical /​ü/​:   = Occasional spellings indicative of sound change (the unrounding in 1.) 3. for historical /​e/​and for historical /​i/​    = Back spellings, which do not indicate that /​e/​and /​i/​have rounded    Rationale: = , = , i.e., they are mere orthographic variants

The second example is taken from the history of English (Stenbrenden 2016: section 5.3), and involves the phonetic developments of the reflexes of Old English (OE) -​ēah and of OE ī. In the first stage, OE -​ēah [æ:əx] (e.g. in high) changed in Middle English (ME), first to something like [əi] or [ei], then to [i:]; [əi~ei] and [i:] seem to have co-​varied for a very long time; the vowel is spelt (and later ), cf. LAEME, LALME and eLALME.1 In other words, are believed to have corresponded to both [ei] and [i:] for some time. Later, the reflex of OE ī (e.g. white) diphthongised to [əi] in the Great Vowel Shift, and there are occasional spellings which seem indicative of this change. However, these for OE ī are almost always dismissed as evidence for the early vowel shift hypothesis, on the grounds that they are back spellings (Kristensson 1987: 65; Stockwell 2006). The logic is that since [i:] in high may be spelt OR , then –​by analogy –​other [i:] (white) may be variously spelt or also. In other words, and may reflect the same phonetic reality, that is [i:], and so are inadmissible as evidence for the diphthongisation of ī. Thus, there is a difference between back spellings and so-​called “occasional” spellings. The motivation behind occasional spellings is often unknown: as the name suggests, they are merely “minor” or “irregular” spellings; in consequence, they are sometimes discarded and sometimes deemed acceptable as evidence of sound-​change. Back spellings, on the other hand, are a subset of occasional spellings, and are clearly the result of false logic; they therefore tend to be rejected point blank as evidence of sound-​change.

1 For the analyses in Stenbrenden (2016), a high number of tokens for words with OE -​ēah were extracted from A Linguistic Atlas of Early Middle English (LAEME, Laing 2008) and A Linguistic Atlas of Late Mediaeval English (LALME, McIntosh –​Samuels –​ Benskin 1986); the latter is now available online in an electronic version (eLALME).

The Evidentiary Status of Back Spellings

197

2. The Status of Back Spellings and the Concept of the “Littera” Traditionally, back spellings have been viewed as the result of phonemic merger: “Reverse or inverse spellings […] always indicate a phonemic coalescence” (Penzl 1957 [1969: 16]). It is for this reason that they are virtually always dismissed as evidence of anything other than the merger that triggered them, which I refer to as the “primary change”,2 or they are treated with extreme caution (Wrenn 1943; McIntosh 1956; Dobson 1968). This view still prevails, albeit mostly implicitly: investigations into the Great Vowel Shift (for example Kristensson 1987: 65) have hesitated to use early for OE /​i:/​as evidence of vowel-​shift diphthongisation, as pointed out previously, because the are said to be back spellings rooted in the merger between the high and white lexical sets.3 However, there is a growing body of research in which irregular spellings are also assessed in respect of the scribe’s orthographic system and of the reliability of the scribe, and potential reasons other than phonemic merger are at least addressed: “The relationship between scribal orthography and historical phonology (figura and potestas) can only be assessed in the context of spelling systems” (Laing 1999: 262; cf. Stenroos 2002). In recent decades, new light has been thrown on the often confusing nature of Middle English spellings and spelling systems with the revived mediaeval concept of the “littera”. The concept was brought to modern scholars’ attention by Benediktsson (1965: 24), but it was Benskin (1982: 20; personal communication) and the group of scholars behind LALME who started using it systematically to explain cases of “confused” letters, back spellings, and the like (LALME III Introduction, 14.3–​14.5; Laing 1998: 279; 1999: 255). A littera (letter), traditionally enclosed in inverted commas, had three attributes: nomen, figura and potestas. The nomen is the name of the letter; the figura(e) the specific letter symbol(s) or shape(s), given in angle brackets; and the potestas/​potestates the sound(s) corresponding to the letter, indicated in square brackets or slanting lines. Capital and lower-​case are examples of two different figurae for the letter ‘a’ (Laing 1999: 255). Importantly, the range of potestates (sound correspondences) is limited for each littera. There is rarely, however, a one-​to-​one relationship in ME scribal systems between letters, their shapes and their sound correspondences. The reason lies 2 It follows that this paper uses the term “secondary change” to refer to secondary changes potentially indicated by back spellings, such as the diphthongisation of ME /​i:/​ in the Great Vowel Shift examined in the preceding discussion. 3 Wells (1982, Chapter 2) established the concept and use of standard lexical sets.

198

Gjertrud F. Stenbrenden

in the situation described in the quote from Laing below, which results in what she (Laing 1999)4 calls “litteral substitution sets”, that is, orthographic or graphemic extensions: the same figura may be used for two or more litterae, and two or more figurae may be used for the same potestas. For example, calligraphic changes in the thirteenth century affected, firstly, the littera ‘þ’, so that it was written without an ascender, “making confusion between the figurae for ‘þ’ and ‘ƿ’ almost inevitable” (Laing 1999: 255); secondly, the littera ‘y’ became lobed, like ‘þ’, and “the figurae for ‘þ’ and ‘y’ in some scripts therefore became identical” (ibid.). Thus, Laing (1999: 256) concludes that a single figura may do service for three litterae and therefore may imply also any of the different possible potestates of those litterae. The logical extension of this phenomenon is the reverse process that many different figurae, or combinations of figurae, may be employed to realise the same potestas.

That is, for the three litterae in question, ‘y’, ‘þ’ and ‘ƿ’, (2) below indicates the figural overlap described above: due to changes in letter shapes, the three letters were all represented by one figura, a -​like symbol, and so could correspond to all the potestates normally attributed to the three letters, i.e. [j θ,ð w] (cf. Benskin 1982: 24). (2)

‘y’ | [j]

‘þ’ | [θ,ð]

‘ƿ’ | [w] 

However, most scribes were exposed to a repertoire of forms (cf. Laing 1998: 280; Laing 1999: 257–​259), so they would at least be familiar with the figurae , , and , in addition to the one they used most commonly, i.e. . Therefore, any of these figurae may be used analogically to represent, say, [j]‌, which is the situation described in the second part of the quote above. For example, young may be spelt in systems in which the figura is commonly used for the litterae ‘y’ and ‘þ’ (Laing 1999: 260; cf. Benskin 1982: 24). In the words of Stenroos (2019: 136), “Once two spellings come to be seen as interchangeable in one context they may be interchangeable in other contexts as well”.

4 Laing (1998: 279) refers to “litteral substitution”, but the concept of substitution sets is not fully developed in this article.

The Evidentiary Status of Back Spellings

199

The concept of “litteral substitution sets” (LSS) has become a household term, but logically, they should be called “litteral or figural substitution sets”: a littera, which is an abstract, rarely merges with another littera. For instance, in those systems where ‘þ’ and ‘y’ are confused (Benskin 1982; see section 4), this confusion is caused by the fact that there is one identical figura for ‘þ’ and ‘y’; this is not to say that the scribes necessarily confused the two abstract litterae –​as long as the figura corresponded to two (or more) different potestates, the scribes would, certainly in principle, be aware of the existence of two letters. As Benskin points out (although using different terminology), the innovative use of substitution sets is hardly the result of ignorant “monoglot Norman copyists” (1982: 20), but the result of someone who in fact knows the language (1982: 21; cf. Laing 1999: 254). In other words, scribes who use litteral/​figural substitution sets do not necessarily lack a littera (e.g. ‘þ’) –​they lack distinctive figurae for two (or more) litterae. Therefore, it is not usually a littera which is substituted for another in a substitution set, but the same figura which is used for different litterae (see further section 4). In the following, I will therefore use the term “litteral/​figural substitution set”, cf. Laing –​Lass (2003: 260, fn. 10), where they admit the need to invoke figural substitution sets at times, but where they state that this “extra degree of complexity is […] not needed for the present discussion”; it seems to be needed here.

3. Alphabetic Writing and the Problem with Back Spellings There is no question that English historical phonology relies on orthographic evidence in the absence of recorded speech (McIntosh 1956: 27): alphabetic writing is phonemic at the outset (Penzl 1957 [1969: 14]), and the underlying principle is that there should be one distinctive letter for each distinctive sound. It therefore seems logical to believe (a) that changes in orthography require an explanation, and (b) that said explanation should involve sound-​change. However, it is equally clear that early spelling, for instance OE orthography, is not phonemic transcript (cf. Stenroos 2002: 449; Stenroos 2019: 135), since sound-​changes may disrupt a previously transparent alphabetic orthographic system. Orthographies may change for other reasons also, for instance as a result of language contact or contact between different spelling conventions (Benskin 1982: 17; Stenroos 2002). Conversely, orthography may standardise and freeze, in which case it becomes “ideographic”: “written ‘words’ tend to become group-​symbols of group-​sounds” (Wrenn 1943: 14; cf. LALME I, section 1.4), and “a group of letters symbolizes collectively a group of noises expressive of an idea or thing” (Wrenn 1943: 14). Thus, after the initial stages of alphabetisation, writing is not always –​perhaps

200

Gjertrud F. Stenbrenden

not even most commonly –​segmental reconstruction of speech; rather, adequate word recognition seems to be the functional criterion. That is, the “function of writing is not to record speech –​it is to communicate linguistic utterances in a different mode from speech” (Stenroos 2002: 453). A person who is a proficient writer “does not recite the words and then transcribe them; instead, he has a ready formula for spelling that he recalls and reproduces” (Stenroos 2002: 454). As a consequence, underlying phonemic systems must instead be inferred from the sounds’ history, from scribal systems, and from analysis of the clashing conventions. This type of exercise is “comparative and not descriptive” (McIntosh 1956: 54): it involves a comparison of the systems of two synchronic stages (Penzl 1957 [1969: 13]) if the true relationship between speech and writing is to be understood in concrete cases. In the words of McIntosh (1956: 28): “only by understanding the limitations of the correlation [between the spoken and written language] can we […] make proper use of the available written material as evidence about the spoken language”.5 With respect to this correlation, there are two basic views, one which assumes that writing simply encodes speech (the “relational” view, cf. Sgall 1987: 2–​3), and another which sees speech and writing as largely independent systems (the “autonomistic” view), but which recognises the link between letters and phonemes (Stenroos 2002: 453). The present writer adopts the latter position: alphabetic writing encodes speech at the outset, and retains the ability to do so, albeit with certain limitations; but the more standardised and conventionalised spelling becomes, the more autonomous the written language. Consequently, at the bottom of the specific problem of back spellings lie two general problems. The first is the correlation between speech and (alphabetic) writing. The other relates to the serious limitations of the Roman alphabet with respect to representing the sounds of Germanic, or of any language that is not Latin. The problem makes itself felt early in English: even in OE, there are cases where one letter is used for two or more sounds (e.g. OE for /​ɡ/​, /​ɣ/​and /​ j/​); later, one sound may be represented by more than one letter (e.g. and for /​k/​). As pointed out previously, conventions may be disrupted by various factors, chiefly sound-​change, language contact and analogy. In this light, occasional spellings may be seen as attempts to test the limits of the Roman alphabet, and thus show interference from speech; if so, irregular spellings (including back spellings) may in certain instances indicate sound-​change.

5 For more on the relationships between the spoken and written medium, see Samuels (1972: 4–​6).

The Evidentiary Status of Back Spellings

201

How do we distinguish back spellings from more general occasional spellings? The digraphs for historical /​i:/​are a case in point: are they to be interpreted as attempts at (phonetic) segmentation (and thus indicate diphthongisation), or are they mere analogical spellings based on the merger between /​i:/​ and /​ei~əi/​in some lexical sets? Penzl (1957 [1969: 14]) states that all back spellings start as occasional spellings, which is true, but a distinction still needs to be made. First, there seems to be an asymmetry in the relationship between such spellings: back spellings all start as occasional spellings, but not all occasional spellings are back spellings. This unremarkable statement merely reflects the general nature of the term “occasional”, i.e. minor spellings which are irregular. The term “back spelling”, however, reflects the fact that such spellings are based on faulty logic. Second, it will be shown later (and it has been hinted at in section 2) that occasional and back spellings, as well as “litteral/​figural substitution sets”, may arise for reasons other than phonemic merger, for instance changes in calligraphy and/​or orthography (Benskin 1982; Laing 1999). Back spellings are thus indicative of “sameness” or common ground at some level, but that level need not be phonemic, but may be phonetic, orthographic, calligraphic or a combination of these. It must be assumed that true back spellings always go both ways. For instance, both -​type and -​type spellings are attested for both the high and white lexical sets, and in some quantity (Stenbrenden 2016, Chapter 5).6 However, the notion of back spellings has been invoked for historical cases in which the alleged back spellings do not go both ways. Stenbrenden (2019b) argues from this fact that phonemic affricates did not exist in OE. Traditional textbooks assume that OE had phonemic affricates, /​dʒ/​(spelt ) in words with palatalised WGmc */​gg/​, and /​tʃ/​(spelt ) in words with a palatalised WGmc */​kk/​. The sole reason why textbooks would assume such a thing is the fact that Sievers (1968 [1895]: 146) pointed out that sequences of etymological [t]‌+[j] (e.g. fetian ‘to fetch’, ort+ġeard ‘orchard’) and [d]+[j] (e.g. mid+ġern ‘fat’) across hiatus are sometimes spelt OE and , respectively, presumably corresponding to [tʃ] and [dʒ] as a consequence of early Yod-​Coalescence (Wells 1982: 207, 247–​248), i.e. /​t+j/​> /​tʃ/​ and /​d+j/​ > /​dʒ/​. Therefore, or so their argument goes, and in and are back spellings, which

6 On the other hand, in high in early ME give way to , which might indicate the chronology of change. That is, by the time of the Great Vowel Shift, words in the high set had /​i:/​and so were discontinued, which meant that the digraphs were “free” to indicate a phonetic diphthong.

202

Gjertrud F. Stenbrenden

are always based on phonemic merger according to the traditional account, and so all OE and must correspond to affricates. However, there are absolutely no -​type spellings for traditional OE , nor -​type forms for OE , in the entire OE corpus; in fact, there are no forms with for OE until early ME, and no forms with for OE until late ME.7 This absence of evidence seems to point to a logical flaw in the argument regarding back spellings, and it raises an important question: in what cases are irregular spellings admissible as evidence of sound change? To answer this question, I will examine further examples of back spellings in the history of English and try to determine their rationale, to see if any patterns emerge.

4. Further Examples of Historical Back Spellings and Their Rationale 4.1 The Change OE eo/​ēo > [ø(:)] Kuhn –​Quirk (1953; 1955) argue, in response to Daunt’s (1939) and Mossé’s (1945) claims to the contrary, that the so-​called “breaking diphthongs” were true phonetic diphthongs. The breaking diphthongs were the result of the fracture of the Pre-​OE front vowels i, e, æ before certain consonants or consonant clusters, i.e. h(+C), lC, rC. Kuhn and Quirk show that in late West-​Saxon (WS), broken and unbroken reflexes of WGmc/​OE e and of WGmc a (> OE æ) are consistently kept apart. Now, the broken reflex of WGmc e is usually spelt in OE, but forms with are also attested, suggesting a rounded vowel, i.e. a phonetic change [eo] > [ø] in late OE. Interestingly, there are also spellings with for etymological OE o, e.g. for worde, for geworhte (Kuhn –​Quirk 1953: 51), but there is nothing to suggest that the reflex of OE o underwent fronting, unrounding, or diphthongisation. In other words, /​o/​never became [ø], [e]‌or [eo], nor merged with the reflex of OE e or eo. Thus, for OE o is a back spelling. The underlying change here is not phonemic merger, but the phonetic change [eo] > [ø]; and the logical basis for the back spellings is simple orthographic extension by analogy. In Laing’s terminology, we have a litteral/​figural substitution set comprising {‘eo’, ‘o’}: the sound [ø(:)] < [e(:)o] could be represented orthographically in different ways, i.e. (etymological, 7 Not even Orm uses a for OE words; he consistently uses the new Caroline always and only for OE (Stenbrenden 2019b: 18). The first in an OE word appears in 1387, cf. the online Middle English Dictionary (Stenbrenden 2019b: 23–​24).

The Evidentiary Status of Back Spellings

203

most common) and (innovative, occasional), the latter reflecting the roundedness of the new vowel. The spelling was used etymologically, however, for the reflex of OE o, and thus OE o could also be spelt by analogy. “The common ground is the spelling . The digraph was thus analogically transferred to words where it did not belong, due to simple graphemic extension, not merger” (Stenbrenden 2016: 149). Penzl’s claim that back spellings always indicate phonemic merger is thus not supported in this instance, but even so, the back spellings for cannot be used as evidence for anything other than the primary phonetic change, OE eo/​ ēo > [ø(:)]. Still, the reverse spellings do seem to provide additional and definite “proof ” of the primary change.

4.2. The Unrounding of OE y/​ȳ The front close rounded phonemes /​y/​and /​y :/​were unrounded in late OE, to [i]‌and [i:], respectively, and subsequently merged with pre-​existing /​i/​and /​ i:/​(Campbell 1959: § 317). In consequence, there are numerous for historical , starting in late OE, but also back spellings with for etymological /​i/​and /​i:/​. Examples are ship and swine (with OE i and ī, respectively), from the later manuscript witness for the Voyage of Ohthere, the eleventh-​century MS Cotton Tiberius B.1 (Bately 1980). In this case, the rationale is indeed phonemic merger between OE /​y/​and /​i/​, and between OE /​y :/​ and /​i:/​, and such spellings are thus not to be taken as evidence of unconditioned rounding of OE /​i/​or /​i:/​;8 rather, and had become co-​variants for /​i/​ and /​i:/​.

4.3. Northern Middle English Scribes and the Use of vs Benskin’s (1982) classic paper examines the use of vs in Northern Middle English (NME). The background is the fact that in NME script, the ascender of had become so short that the figurae of the litterae ‘y’ and ‘þ’ had become identical,9 as a mostly -​like symbol (Benskin 1982: 13, 22, 24). When scribes from the North or the North-​East Midlands attempted a southern system, they thus had to re-​learn the symbol and to use it correctly, etymologically speaking. As is so often the case when language users try to acquire a 8 However, OE /​i/​and /​i:/​do seem to have rounded to /​y/​and /​y :/​in certain phonetic environments (Campbell 1959: §§ 315, 318). 9 For the same reason, the figura for ‘þ’ (wynn) was also indistinguishable from (Benskin 1982: 24).

204

Gjertrud F. Stenbrenden

distinction which is absent from their own system, hyper-​corrections result, e.g. young, will (Benskin 1982: 14, 17) Thus, young is a back spelling, but the rationale is purely calligraphic: these spellings certainly do not indicate that [j]‌= [θ], as there is no potestatic overlap between the two litterae. Hence, we have a litteral/​figural substitution set involving {‘y’, ‘þ’}, and the common ground is the figura . The motivation behind the back spellings is thus not phonemic merger; in fact, not even phonetic change. The back spellings therefore cannot be taken as evidence for anything other than a change in script, whereby the shape of ‘thorn’ had become identical to the shape of ‘y’, and that the other figura, , was consequently confused with in scribal attempts to imitate a system which maintained a distinction between the two.

4.4. Litteral/​Figural Substitution Sets involving ‘y’, ‘þ’, ‘ƿ’ and ‘ʒ’ Benskin (1982: 20–​21) comments on the use of for , which hinges on the fact that ‘þ’ and ‘y’ were not distinguished in terms of figurae: could thus correspond to [j]‌in certain words, and so could . Thus, if year could be spelt or , there was no reason why that could not be spelt . The latter is a back spelling and the rationale is originally “purely calligraphic” (Benskin 1982: 21), though the common ground between and is the sound [j].10 The litteral/​figural substitution set comprises {‘y’, ‘þ’, ‘ʒ’}, in which ‘y’ = ‘þ’ (figura), and ‘y’ = ‘ʒ’ (potestas), therefore ‘þ’ = ‘ʒ’ (analogy).11 Laing (1999) elaborates on complex scribal systems in which not just ‘y’ and ‘þ’, but also ‘ƿ’ and ‘ʒ’, are used interchangeably (see also Benskin 1982: 16). This situation came about through a series of calligraphic changes, described in sections 2 and 4.3 above. To reiterate, the figura for ‘þ’ lost its ascender and became identical to (hence, ‘thorn’ and ‘wynn’ became interchangeable); then became lobed, like , and ‘y’ subsequently fell in with ‘thorn/​wynn’ (cf. Benskin 1982: 24); that is, there was one figura for all three litterae, and later also for ‘ʒ’ (‘yogh’) for the reason outlined in the preceding paragraph. Here there was also some potestatic overlap: [j]‌had traditionally been represented by or , and now and could also –​by analogy –​be used for the same

10 The common ground is [j]‌because ‘y’ and ‘ʒ’ could both historically correspond to [j], so when the figurae for ‘y’ and ‘þ’ became indistinguishable, ‘þ’ could also correspond to [j]. In the next stage, when ‘ʒ’ joined the set, the litterae ‘y’, ‘þ’ and ‘ʒ’ could be used interchangeably for all their combined potestates. 11 See also CoNE, which treats this and similar phenomena under the heading “orthographic remapping”: .

The Evidentiary Status of Back Spellings

205

sound. Hence, and for [j] are back spellings (e.g. young and give, adduced by Laing 1999: 260). In this case, therefore, we have a litteral/​figural substitution set involving {‘y’, ‘þ’, ‘ƿ’, ‘ʒ’}, and the rationale is mostly calligraphic, but also orthographic and phonetic. What Laing shows, above all, is that complex or “prodigal” ME scribal orthographies are not unsystematic or asystematic (Laing 1999: 261), despite the apparent orthographic confusion: there is in fact a discernible system, which has come about as a result of calligraphic changes and orthographic extension (usually by analogy).

4.5. Final Devoicing or Neutralisation of /​d/​ Laing –​Lass (2003) investigate back spellings which seem to indicate final devoicing of /​d/​or even final neutralisation between /​t/​and /​d/​. The spellings in question are the co-​variants and good, which in turn lead to co-​variants and that; these latter are the back spellings. The underlying change appears to be [-​d] > [-​t], i.e. devoicing of /​d/​in final position, so that ‘d’ = ‘t’ in final position; this leads, however, to the faulty assumption12 that ‘t’ = ‘d’ in final position, and thus we have a litteral/​figural substitution set involving {‘d’, ‘t’} in this particular phonotactic position. Hence, phonetic change is at the bottom of the back spellings, but distributional phonology also plays a role, as does the spurious conclusion on the scribes’ part that two litterae may be used interchangeably in final position. The back spellings in this case may be used as evidence of the final devoicing of /​d/​, but not of any voicing of /​t/​in the same position.

4.6. Middle English Epenthetic LALME and eLALME attest a few spellings of the type kind, which may be usefully compared to e.g. might and high, in which is historical and reflects earlier /​x/​. The reason for the use of in kind seems to be that the sound etymologically corresponding to , i.e. /​x/​, had vocalised completely in might and high; in such words, therefore, was a mere historical remnant, but it usefully indicated that the preceding vowel was

12 The assumption is faulty because the context-​dependant equation ‘d’ = ‘t’ is rooted in a phonetic change, i.e. devoicing, which did take place in final position; but the opposite process, voicing of plosives, did not happen in this position, nor has ever been common in English.

206

Gjertrud F. Stenbrenden

long, or had become long. The logic behind in kind therefore seems to be that is used to indicate vowel length in words which had never had postvocalic /​x/​. Interestingly, is rarely used as a length diacritic in words whose vowel has always been long (e.g. white), but such usage is indeed recorded in eLALME for while, fire, life. The question is whether and are members of a ‘litteral/​figural substitution set’. Similar usage of is found for other vowels also, e.g. for OE /​u:/​in cases where the vowel was not historically followed by /​x/​. LAEME #282 (Ely, 1275–​1325) records , thou, out, and clout; the form how is found in Trevisa’s translation of Higden’s Polychronicon (Sisam 1962: 146, l. 13); LALME has numerous forms, e.g. now in Ely 673, out in Li 210, out in Nfk 642.13 These and similar forms could be back spellings resting on the vocalisation of [x]‌in words such as ought. There is also a possibility that the sequence indicates a minimal diphthong with a tense off-​glide. Whatever the answer is to the question posed in the preceding, the rationale seems phonemic: in sequences of etymological high vowels + /​x/​, the postvocalic fricative had vocalised and the preceding vowels were, or had become, long; the new long vowels merged with pre-​existing long vowels. As spelling is conventional, the was retained and could thus be used to indicate the length of the preceding vowel, even in words in which the vowel had never been followed by /​x/​. In other words, in kind may probably be taken as evidence of a long vowel, but certainly not as evidence of an epenthetic fricative.

4.7. Back Spellings Involving OE e/​ē and eo/​ēo Stenroos (2019: 133, 136) addresses back spellings explicitly, i.e. fourteenth-​and fifteenth-​century back spellings for OE e/​ē and eo/​ēo (which are of a different kind than those discussed in section 4.1). Briefly, the reflexes of OE eo/​ēo were monophthongised to front rounded [ø, ø:] in late OE, then unrounded to [e, e:], merging with the reflexes of pre-​existing e/​ē. The unrounding took place in the twelfth century in the North, the East and the South Midlands, but not until the fourteenth century in the West Midlands (Wright –​Wright 1928: §§ 58–​ 67). Thus, in the West Midlands, the two sets of vowels are often kept apart orthographically in ME, -​type spellings being the norm for OE e/​ē, and -​type14 spellings co-​varying with for OE eo/​ēo. Frequent forms with for the reflexes of OE eo/​ēo in the West Midlands of course suggest that unrounding took place there too, but spellings indicative of a retained rounded vowel are attested from the same area in the late ME period (Stenbrenden 2016: 119). In other words, unrounded and rounded reflexes of OE eo/​ēo must have existed side by side in the West and South-​West for a considerable length of time. Stenroos considers spellings for OE eo/​ēo in late ME texts of various genres in the MEG-​C corpus. Of almost 17,000 tokens for OE eo/​ēo, only about 1,000 may be taken to indicate a retained front rounded vowel, in texts localised to an area that “might be called a slightly extended version of the ‘traditional South-​West Midland area’ ” (Stenroos 2019: 141). More importantly, there are 142 unhistorical cases of , i.e. back spellings, which interestingly are found in texts from the same area, and which mostly fall into three groups: is used for (a) vowel after , (b) words with OE y/​ȳ, and (c) unstressed vowels (Stenroos 2019: 150). Stenroos suggests that therefore is part of a substitution set which includes the figurae and and which generally indicates roundedness of the vowel. This substitution set hinges on the fact that retained [y, y:] was spelt in the West Midlands, and that there are “crossovers of spelling between Old English eo/​ēo and y/​ȳ words” in early ME (Stenroos 2019: 150); in fact, the two front rounded vowels may even have merged (Sundby 1963; Ek 1972). Thus, both and corresponded to rounded vowels. In an earlier study, Lass –​Laing (2005) conclude from analysis of early ME spellings that the sound distinction between the reflexes of OE eo/​ēo and those of OE e/​ē was in fact not upheld in ME, their conclusion being based on the difficulty in mapping to any specific etymological category and on litteral substitution sets.15 Stenroos, however, points out that such substitution sets “can be problematic as explanations as they are in principle endless” (2019: 137); i.e., there is no limit as to the number of litterae in a litteral/​figural substitution set. While this is true, it does not, in the present writer’s opinion, invalidate the concept or indeed the rationale behind back spellings: back spellings are analogical 14 -​type spellings include at least the following spellings: , , , , , . 15 Lass –​Laing’s (2005) arguments apply to /​y/​and /​y :/​as well. Their claim that front rounded vowels were not retained in the West Midlands in ME is partly refuted by Stenbrenden (2016: 273): “the SW Midlands spellings for OE ȳ and OE ū in LAEME imply merger of figurae, but not of potestates, as otherwise it would have been impossible for the reflexes of the two sounds to unmerge later”, except as a result of contact between two different phonemic systems.

208

Gjertrud F. Stenbrenden

extensions based on the faulty logic that if “two spellings come to be seen as interchangeable in one context they may be interchangeable in other contexts as well” (Stenroos 2019: 136), which is the case for the unhistorical adduced in the preceding. This is indeed recognised by Stenroos (2019: 151): “regular sound-​ spelling correspondences need not be biunique, and the extension of spelling units to other functions does not cancel earlier mappings” between sounds and spellings. Like the present writer, Stenroos (2019: 152) is forced to discard the traditional view that phonemic merger is always behind back spellings: “other evidence suggests quite strongly that Middle English spelling simply does not work like this” and analogical spellings “could take place simply as a matter of letter substitution”.

4.8. Summary It is clear that back spellings abound in the history of the English language, but their motivation varies considerably, and the underlying change in not always phonemic merger, despite Penzl (1957 [1969]). In fact, their rationale is most commonly not phonemic, but rooted in analogy and changes in calligraphy and orthography. That is not to say, however, that phonemic merger and phonetic changes do not cause back spellings; they obviously do, and more often than not, back spellings arise as a result of a complex interplay between analogy and the various levels of language. There is always a “smallest common denominator” or a reason for the analogy, and this denominator tends to be figural and/​or potestatic overlap between historically distinct categories.

5. The Relationship Between Writing and Speech, Again At this stage, it might be useful to re-​visit the problem addressed in section 3 and to ask the question: What resources are available to indicate sound-​change? As pointed out in the preceding, alphabetic writing is phonemic at the outset: there is one distinctive letter for each distinctive sound. However, sound-​changes may later upset this one-​to-​one system, and spelling becomes largely, but not uniquely, ideographic (in the sense reported previously). Besides, alphabets devised for specific languages are far from ideal for other languages: the Roman alphabet was severely limited when it came to represent the sounds of Gmc, in particular front rounded vowels, diphthongs, velar fricatives, phonemic splits among the consonants, etc. (cf. Daunt 1952: 52). The point to be made is that simply assuming that spelling is (always) ideographic overlooks the fact that alphabets do have a phonemic basis and are

The Evidentiary Status of Back Spellings

209

suitable for representing sounds: “spellings do have a real relationship to the sound values they conventionally imply” (Laing 1999: 251). Also, the spoken and written media “must keep in step, for, if the level of correspondences between them drops, the written language becomes an uneconomic medium” (Samuels 1972: 5). It also overlooks phonetic salience: given that sound-​change has occurred, and that the results of this change are salient enough to be noted by scribes (that is, that there is a discrepancy between writing and speech), the scribes have to be resourceful with the Roman alphabet if they wish to indicate the change by orthographic means. In the simplest terms, their attempts at sound segmentation are what we call “occasional” or “irregular” spellings (Penzl 1957 [1969: 14]). With this in mind, we can assess the early ME forms again. When OE/​eME /​i:/​diphthongised in the Great Vowel Shift, the first stages were probably [ɪi] > [əi], if similar processes in Present-​Day English (PDE) are anything to go by (Wells 1982, section 3.4.2); the last stage would have been [ʌi] or [ai]. The diphthongs [əi] and [ʌi] are salient, and markedly different from [i:], so the question is how they could be represented by a scribe using the Roman alphabet. In the absence of letters for central vowels, it is fair to assume that he would have used , later –​which is exactly the occasional spellings we find for the reflex of early ME /​i:/​in ME and eModE.16 In this context, it may be useful to consider the “no merger with X argument”, which is often encountered in attempts to refute the early vowel shift hypothesis. There is no dearth of ME spellings which seem to indicate that the Great Vowel Shift started much earlier than the conventional date given in textbooks, which is c. 1400. Stenbrenden (2016) adduces a high number of such spellings from c. 1250 onwards, for the reflexes of OE /​i:/​, /​u:/​, /​e:/​and /​o:/​. However –​as pointed

16 Notably, in the similar process of diphthongisation of PDE /​i:/​, for instance in the Diphthong Shift of Cockney and popular London speech (Wells 1982, section 3.4.2), the resulting diphthong (phonetically [əi]) is usually rendered as , because conventionally correspond to /​eɪ/​in PDE and is the closest orthographic and phonetic fit. The blog at http://​dialectblog.com/​british-​accents/​ provides a good modern example, stating the following under the headline “London vowel shift”: “The vowel sounds are shifted around so that Cockney ‘day’ sounds is [sic] pronounced IPA dæɪ (close to American ‘die’) and Cockney buy verges near IPA bɒɪ (close to American ‘boy’)”. That is, modern “scribes” attempt to provide phonetic segmentation to indicate sound-​changes, exploiting the conventional correspondences, in PDE, between certain spellings and their usual pronunciations. The author of the blog, unlike mediaeval scribes, makes the inaccuracies clear by using the phrase “close to”.

210

Gjertrud F. Stenbrenden

out before –​the for OE /​i:/​are dismissed as evidence for a phonetic diphthong for two reasons: (a) because they might be back spellings caused by the merger between the high and white lexical sets, and (b) because the reflex of OE /​i:/​never merged with the reflex of historical ei (Dobson 1968: 660–​663); curiously, late ME and early ModE forms with tend to be accepted as evidence of advanced diphthongisation, although the argument in (b) applies equally here (Dobson 1968: 663, note 4). Similarly, for the reflex of OE /​u:/​is often not accepted as evidence of a phonetic diphthong (Wyld 1914: 182) for two reasons: (a) the etymological diphthong au had monophthongised to [ɔ:] at this stage (late fifteenth century),17 and (b) the reflex of OE /​u:/​never merged with the reflex of historical au. Using the same line of argument, early for OE /​e:/​and early for OE /​o:/​could be similarly dismissed, because these vowels did not merge with pre-​existing /​i:/​and /​u:/​, respectively. But this line of argumentation loses sight of the fundamental rationale of alphabets, i.e. that they are sound-​based: alphabetic spelling may be both ideographic and sound-​sensitive at the same time. Besides, regarding the diphthongisation of the two early ME close long monophthongs /​i:/​and /​u:/​, how is a scribe to indicate a phonetic diphthong other than by putting two vowel letters together? It is to be expected that he would choose the two vowel letters that most closely match the start and end points of the diphthong, respectively. That would mean and for the diphthongised reflex of ME /​i:/​, as stated in the preceding, and , and for the reflex of ME /​u:/​, depending on how far the first element of the diphthong had dropped in articulatory space. Again, all of these spellings are attested for the reflex of ME /​u:/​in the surviving ME material. To my mind it is clear that the digraphs in question are indeed attempts at phonetic or segmental analysis, and such a notion is at least acknowledged by Dobson when he states that “the transcription ei [for [əi] < /​i:/​] was only approximate” (1968: 660); that is, the orthoepists attempted to indicate [əi] and the closest orthographic match was (which was also the conventional digraph corresponding to the etymological ei diphthong).

17 Wyld finds it odd that the digraph should be used to indicate a phonetic diphthong when etymological au had monophthongised, but this is exactly the problem with such arguments: the development of etymological diphthongs into monophthongs is taken as evidence that a digraph cannot be used to indicate a phonetic diphthong. In later work, Wyld (1936: 230–​232) seems more open to the idea that such spellings indicate diphthongisation of ME /​u:/​.

The Evidentiary Status of Back Spellings

211

6. The Way Forward To determine the evidentiary status of back spellings, we must ask questions of the text in order to (1) work out the rationale and the analogy behind the back spellings, (2) assess the role of the scribe and lay out his system, and possibly (3) identify clashing systems. (1) Work out the rationale and the analogy As I have shown in the preceding, the logic behind back spellings varies immensely. Thus, calling a spelling “analogical” need not entail its rejection as evidence of sound-​change –​the true nature of the analogy and the underlying linguistic change must be worked out before the evidentiary status of the back spelling may be settled. If phonemic merger is indeed the reason, the back spelling may or may not indicate a second sound-​change and must therefore be treated with caution. It is precisely in such cases that the scribe’s entire system must be submitted to close analysis (more on this below). Establishing litteral/​ figural substitution sets should also be attempted. If no such substitution sets can be established, the likelihood increases that the spelling in question is phonetically motivated. However, as we have seen, the back spelling will in most cases be rooted in a substitution set of some kind. If the substitution set hinges on purely calligraphic changes (as in the case of ‘þ’ and ‘y’), there is no reason to suppose an underlying sound-​change; the same is in fact usually true if there is (also) potestatic overlap. For instance, there was potestatic overlap in ME between ‘y’ and ‘ʒ’ for [j]‌; as ‘y’ was also involved in a litteral/​figural substitution set with ‘þ’ and ‘ƿ’ (for calligraphic reasons), ‘ʒ’ could also be included in the same set (because ‘y’ = ‘ʒ’) and could thus be used in the same contexts as ‘þ’ and ‘ƿ’ (because ‘y’ = ‘þ’ = ‘ƿ’). Lastly, it must be determined whether the back spellings go both ways; if not, the likely rationale is probably not phonemic merger (cf. Stenroos 2002: 457). (2) Assess the role of the scribe Given the somewhat fuzzy nature of substitution sets, back spellings should be thoroughly assessed in terms of the scribal system in which they appear; thus, it must be determined whether the scribe is a literatim copyist or translator (Benskin –​Laing 1981), if he has a conservative or innovative approach to spelling, and if he uses other occasional or irregular spellings, and of what type. For instance, if the scribe consistently uses irregular spellings that might indicate vowel shift, and he does so for more than one vowel, the case is stronger for viewing these spellings as indicative of sound-​change. Irregular digraphs may be

212

Gjertrud F. Stenbrenden

thought to correspond to diphthongs if the scribe spells etymological diphthongs consistently and thus shows that he is aware of the nature of diphthongs. It is perfectly possible to systematically analyse scribal systems and to draw likely, even probable, inferences about sound-​change from such analyses (see, for instance, Laing –​Lass 2003 and Stenbrenden 2019a, using the tagged texts in LAEME). A case in point is the scribal text given as #278 in LAEME (MS London, British Library, Cotton Caligula A.ix, Laʒamon A, Hand B), which has been localised to Worcestershire and dated to the period 1250–​1275. It has numerous for the reflex of OE /​u:/​(in e.g. about, house, how, thou), possibly indicating vowel-​ shift diphthongisation; but since is also used for the vowel in self (< OE self, seolf) which is also recorded with , the forms with for historical /​u:/​may be back spellings. However, a trigraph is also attested in how and thou as well as in you (< OE ēower) and four (< OE fēower), which strengthens the case for interpreting these irregular spellings as corresponding to a diphthong. Interestingly, the same scribal language also features and good, as well as here and flee imp., which point to vowel-​shift raising of the reflexes of OE /​o:/​and /​e:/​, respectively. Thus, a detailed inspection of the scribe’s entire system, including back spellings, reveals a truly idiosyncratic orthography which is suggestive of vowel shift around the mid-​thirteenth century. (3) Identify the clashing systems Stenroos (2002: 458–​461) suggests a procedure for establishing what systems lie behind orthographic variation, especially in cases where authorial and scribal systems clash. As an example, she uses spellings for the OE short low and mid unrounded vowels a, æ, ea and e and lays out the various systems that may be thought to lie behind the actual manuscript spellings in Laʒamon A, Hand A (MS London, British Library, Cotton Caligula A.ix); her Table 2 (Stenroos 2002: 466) is reproduced (with minor changes) in Figure 1 below. System 1 OE spelling (presumed phonemic) a æ ea e

System 2 Author/exemplar spoken system

System 3 Author/exemplar written system

/a/ /æ/

a æ (ea)

/e/

e

System 4 Scribe’s own system a

Output MS a a(æ)((e,ea))

e

e((æ))

Figure 1:  The interplay between scribal and authorial systems in Laʒamon A, Hand A (taken from Stenroos 2002: 466). The use of (double) parentheses in this figure follows LALME practice: single round brackets indicate a secondary variant, whereas double round brackets indicate a minor variant (McIntosh et al. 1986).

The Evidentiary Status of Back Spellings

213

Starting with the MS output to the far right, it is clear that the reflexes of OE a and e are consistently spelt and , respectively, by the Caligula scribe, whereas spellings for OE æ and ea are variable (Stenroos 2002: 451); besides, closer analysis reveals that the three open(-​mid) unrounded vowels seem to have merged under a in the scribe’s own system, whereas e remains distinct (despite the minor for OE e). The scribe is however copying an exemplar which appears to show a three-​way contrast, reflecting the exemplar scribe’s (or author’s) written and spoken system, in which OE e and a are retained, and in which the reflexes of OE æ and ea have merged under æ ; the minor tokens could thus be back spellings reflecting a phonemic merger between etymologically separate sounds, and are conditioned by the written system of the exemplar as well as by the Caligula scribe’s repertoire of litterae from Latin, French and English usage (Stenroos 2002: 463–​465). The relational view of writing, suggesting that writing merely encodes speech, falls short of explaining this orthographic variation, whereas Stenroos’s approach satisfactorily accounts for the variation as a result of clashing systems. In concrete terms and with ME phonology in mind, the analysis reveals that in all likelihood, the writer of the exemplar had a system in which there were three short low or mid unrounded vowels (something like /​e/​, /​æ/​, /​a/​), whereas the Laʒamon scribe had a two-​ way contrast (probably between /​e/​and /​a/​) in the late thirteenth century. Thus, despite the shortcomings of spelling as evidence of sound-​change, it is possible to uncover ME phonologies through fine-​grained analysis of regular, irregular and back spellings in a given text language, and of the interaction between the written and spoken systems of the text scribe and those of his exemplar.

7. Conclusions It is clear that the logic or motivation behind back spellings varies greatly, and this fact should be properly recognised; the reasons for back spellings of course always include analogy, in combination with changes on the levels of calligraphy, orthography, phonetics, and phonology. Consequently, calling a spelling “analogical” need not entail the prima facie rejection of it as evidence of sound-​ change, at least not until the precise nature of this analogy has been established. Additionally, a thorough assessment of scribal approaches and systems is likely to reveal whether back spellings indicate sound-​change in certain cases. To re-​ write an old adage: “each back spelling has its own history”. Stenroos (2002: 458) calls for a theory of orthographic variation and change. Although she does not present a theory as such, she does suggest a procedure for identifying clashes between authorial and scribal systems, both written and

214

Gjertrud F. Stenbrenden

spoken, and demonstrates that such a procedure may enable the linguist to identify the phonological systems behind variable spellings. It is the present writer’s opinion that this method may be usefully applied to cases of back spellings also. We need a systematic approach, asking questions of the text in their capacity as witnesses; determining whether a spelling is occasional or analogical, as well as its evidentiary status, depends on the answers to these questions.

References Alcorn, Rhona –​Bettelou Los –​Joanna Kopaczyk –​Benjamin Molineaux (eds.). 2019. Historical Dialectology in the Digital Age. Edinburgh: Edinburgh University Press. Bately, Janet. 1980. The Old English Orosius. EETS SS 6. London: Oxford University Press. Benediktsson, Hreinn. 1965. Early Icelandic Script. Reykjavík: The Manuscript Institute of Iceland. Benskin, Michael. 1982. “The letters and in later Middle English, and some related matters”. The Journal of the Society of Archivists 7.1: 13–​30. Benskin, Michael –​Margaret Laing. 1981. “Translations and Mischsprachen in Middle English manuscripts”. In: Benskin, Michael –​Michael L. Samuels (eds.): 55–​106. Benskin, Michael –​Michael L. Samuels (eds.). 1981. So meny people longages and tonges. Edinburgh: Middle English Dialect Project. Campbell, Alastair. 1959. Old English Grammar. Oxford: Clarendon. Daunt, Marjorie. 1939. “Old English Sound-​ Changes Reconsidered in Relation to Scribal Tradition and Practice”. Transactions of the Philological Society: 108–​137. Daunt, Marjorie. 1952. “Some Notes on Old English Phonology. A reply to Mr. M. Samuels’ paper”. Transactions of the Philological Society 51.1: 48–​54. Dobson, Erik J. 1968. English Pronunciation 1500–​ 1700. Vol. I. 2nd edn. Oxford: Clarendon Press. Ek, Karl-​Gustav. 1972. The Development of OE ȳ and ēo in South-​Eastern Middle English. Lund: Gleerup. eLALME = Benskin, Michael –​Margaret Laing –​Vasilis Karaiskos –​Keith Williamson (eds.). 2013–​. An electronic version of a Linguistic Atlas of Late Mediaeval English. http://​www.lel.ed.ac.uk/​ihd/​elalme/​elalme.html [accessed 4 September 2020]. Kristensson, Gillis. 1987. A Survey of English Dialects 1290–​1350: the West Midland Counties. Lund: Lund University Press.

The Evidentiary Status of Back Spellings

215

Kuhn, Sherman M. –​Randolph Quirk. 1953. “Some Recent Interpretations of the Old English Digraph Spellings”. Language 29: 143–​156. Kuhn, Sherman M. –​Randolph Quirk. 1955. “The Old English Digraphs: A Reply”. Language 31.3: 390–​401. LAEME = Laing, Margaret. 2008. A Linguistic Atlas of Early Middle English. Compiled by Margaret Laing with accompanying software by Keith Williamson. University of Edinburgh. www.lel.ed.ac.uk/​ihd/​laeme2/​laeme2. html [accessed 4 September 2020]. Laing, Margaret. 1998. “Raising a Stink in The Owl and the Nightingale: a New Reading at line 115”. Notes and Queries 243: 276–​284. Laing, Margaret. 1999. “Confusion wrs confounded: litteral substitution sets in Early Middle English writing systems”. Neuphilologische Mitteilungen 100.3: 251–​270. Laing, Margaret –​Roger Lass. 2003. “Tales of the 1001 nists: the phonological implications of litteral substitution sets in 13th-​century South-​West-​Midland texts”. English Language and Linguistics 7.2: 1–​22. Lass, Roger (ed.). 1969. Approaches to English Historical Linguistics: an anthology. New York: Holt, Rinehart and Winston. Lass, Roger –​Margaret Laing. 2005. “Are Front Rounded Vowels Retained in West Midland Middle English?” In: Ritt, Nicolaus –​Herbert Schendl (eds.): 280–​290. Luelsdorff, Philip A. (ed.). 1987. Orthography and Phonology. Amsterdam/​ Philadelphia: John Benjamins. McIntosh, Angus. 1956. “The Analysis of Written Middle English”. Transactions of the Philological Society: 26–​55. McIntosh, Angus –​Michael Samuels –​Michael Benskin. 1986. A Linguistic Atlas of Late Mediaeval English, I–​IV. Aberdeen: Aberdeen University Press. Mossé, Fernand. 1945. Manuel de l’anglais du Moyen-​ Âge, I. Vieil-​ anglais. Paris: Aubier. Penzl, Herbert. 1957 [1969]. “The Evidence for Phonemic Changes”. In: Pulgram, Ernst (ed.): 193–​208. Reprinted in Lass, Roger (ed.): 10–​24. Pulgram, Ernst (ed.). 1957. Studies Presented to Joshua Whatmough on his Sixtieth Birthday. The Hague: Mouton. Ritt, Nicolaus –​Herbert Schendl (eds.). 2005. Rethinking Middle English: Linguistic and Literary Approaches. Frankfurt am Main: Peter Lang. Samuels, Michael L. 1952. “The study of Old English phonology”. Transactions of the Philological Society 51.1: 15–​47.

216

Gjertrud F. Stenbrenden

Samuels, Michael L. 1972. Linguistic Evolution, with special reference to English. Cambridge: Cambridge University Press. Schaefer, Ursula. (ed.), The Beginnings of Standardization: Language and Culture in Fourteenth-​Century England. Frankfurt am Main: Peter Lang. Sgall, Petr. 1987. “Towards a Theory of Phonemic Orthography”. In: Luelsdorff, Philip A. (ed.): 1–​30. Sievers, Eduard. 1968. An Old English Grammar, translated and edited by Albert S. Cook. New York: Greenwood Press. Stenbrenden, Gjertrud F. 2011. The Chronology and Regional Spread of Long-​ Vowel Changes in English, c. 1150–​1500. PhD dissertation, University of Oslo. Stenbrenden, Gjertrud F. 2016. Long-​Vowel Shifts in English, c. 1050–​1700. Cambridge: CUP. Stenbrenden, Gjertrud F. 2019a. “The development of OE ǣ: Middle English spelling evidence”. In: Alcorn, Rhona –​Bettelou Los –​Joanna Kopaczyk –​ Benjamin Molineaux (eds.): 113–​132. Stenbrenden, Gjertrud F. 2019b. “Old English and its sound correspondences in Old English and Middle English”. English Language and Linguistics: 1–​32. DOI: https://​doi.org/​10.1017/​S1360674319000182. Stenroos, Merja. 2002. “Free Variation and Other Myths: Interpreting Historical English Spelling”. Studia Anglica Posnaniensia 38: 445–​468. Stenroos, Merja. 2019. “The Development of Old English eo/​ēo and the Systematicity of Middle English Spelling”. In: Alcorn, Rhona –​Bettelou Los –​ Joanna Kopaczyk –​Benjamin Molineaux (eds.): 133–​155. Stockwell, R.P. 2006. “The Status of Late Middle English Spellings as Early Evidence of the English Vowel Shift”. In: Schaefer, Ursula (ed.): 175–​180. Sundby, Bertil. 1963. Studies in the Middle English Dialect Material of Worcestershire Records. Bergen: Norwegian Universities Press. Wells, John C. 1982. Accents of English, Vol. 1. Cambridge: Cambridge University Press. Wrenn, C.L. 1943. “The Value of Spelling as Evidence”. Transactions of the Philological Society 42.1: 14–​39. Wright, Joseph –​Elizabeth M. Wright. 1928. An Elementary Middle English Grammar. Oxford: Oxford University Press. Wyld, Henry C. 1914. A Short History of English. London: John Murray. Wyld, Henry C. 1936. A History of Modern Colloquial English. Oxford: Basil Blackwell.

Merja Stenroos

What, If Anything, Are Middle English Dialects? Some Thoughts on a Changing Concept Abstract: Middle English is traditionally thought of as the “dialectal phase” of English: the period when dialectal variation is reflected in the written record. There is general agreement that the written variation is to a large extent geographically conditioned: however, what precisely it reflects and how it relates to the physical map are remarkably controversial questions. Approaches to dialectal variation have changed dramatically over the last century, as have attitudes to written language; however, the implications of such changes may take a long time to permeate scholarly traditions. This chapter reviews some of these major changes and discusses their implications for the study of Middle English variation. It argues that there can be no single absolute relationship between linguistic form and geography: rather, there are numerous ways of relating the two, and consequently innumerable different dialect maps. Middle English dialects, accordingly, emerge from the data: whether as constellations shaped by the questions we ask or as observational artifacts based on our subjective experience. Keywords: Middle English, dialects, dialect areas, dialect continuum, social space, linguistic geography

1. Introduction One of the best known statements about Middle English, at least to linguists, is Barbara Strang’s (1970: 225) point that “[d]‌ialectology is more central to the study of ME than to any other branch of English historical linguistics”. In his chapter in the Cambridge History of the English Language, Milroy (1992: 160) added to Strang’s statement that “virtually every piece of research into Middle English language [...] is, at least potentially, a contribution to Middle English dialectology”. Such statements reflect the fact that surviving Middle English texts show an exceptional amount of variation in the written mode, compared to most writing systems in any period. Because of this variation, usually considered to reflect geography at least to some extent, Middle English is often referred to as the “dialectal” period of English, preceded and succeeded by periods of relative homogeneity in the written record. Much of the past scholarship dealing with linguistic

218

Merja Stenroos

variation in Middle English has situated itself under the heading “Middle English dialectology”, and virtually any textbook on the history of English contains a section on Middle English dialects, with or without a map. At the same time, there is little consensus on how these dialects are defined: the term “Middle English dialect” may refer to rather different concepts, based on different models of linguistic and geographical variation, and with important implications for how we deal with the evidence and even how we think about the history of written English. The differences stem partly from changing models of linguistic variation, partly from the challenges posed by different kinds of evidence. An (anonymized) example is in order here. At a conference on historical English some years ago, a plenarist made a point in passing that when it comes to regional variation in Middle English texts, it does not matter where a text comes from: the important factor is not the geography but the dialect. This point was clearly meant to be self-​evident for at least part of the audience, and it does indeed make sense in a very specific research tradition: that of the Linguistic Atlas of Late Mediaeval English (henceforth LALME); seen from the point of view of other approaches to dialectology, however, such a statement might come across as contradictory, even nonsensical. The idea that geography and (regional) dialect may be contrasted reflects a particular set of assumptions about linguistic variation in written Middle English, which is not necessarily universal; it is clear that scholars may mean quite different things when referring to Middle English dialects. It is, of course, unavoidable, even important, that different research traditions use terms differently, and base their work on different models. The contention of the present chapter is, however, that such differences may be particularly problematic in the study of past linguistic variation, as the challenges posed by the incomplete evidence mean not only that scholars tend to build much on earlier work, but also that our basic assumptions about variation may shape the findings to a larger extent than would be the case with contemporary language. The purpose here is to consider some of the main approaches to Middle English dialects over the last century: what kind of models of variation they relate to, and what, if anything, they might have to do with geography. The intention is not to promote a single definition of Middle English dialects; rather, the usefulness of such general definitions is called into question. At the same time, as any work on Middle English linguistic variation will build on a foundation laid by earlier scholars, with varying ideas about the geographical patterning of language, the question addressed is to what extent we can assume the different frameworks to be compatible at all.

What, If Anything, Are Middle English Dialects?

219

2. The Concept of Dialect The Oxford English Dictionary Online gives two main senses of the term dialect that are relevant for the present purpose: 2. A form or variety of a language which is peculiar to a specific region, esp. one which differs from the standard or literary form of the language in respect of vocabulary, pronunciation, idiom, etc. (as a mass noun) provincial or rustic speech. 3. Manner of speaking, language, speech; esp. the mode of speech peculiar to, or characteristic of, a particular person or group; phraseology, idiom; jargon; a particular variety of any of these.

Definition (2) sees dialect as a variety of language that relates to geography, as well as being different from any standard or literary form. On the other hand, definition (3) suggests that a dialect may be any kind of variety, not necessarily one based on geography, nor indeed a non-​standard one. Both senses of the term appear commonly in the literature, whether or not they are made explicit. In their classic work titled Dialectology, Chambers and Trudgill (1980 [1997: 3]) list several definitions of dialect which they do not themselves endorse: a) a “substandard, low-​status, often rustic form of language” associated with groups lacking in prestige, b) a language which has no written form, and c) deviations from a norm, “aberrations of a correct or standard form of language”. All these senses of the term, which have been mainly used by non-​linguists, contrast dialects in one way or another with a “proper” or “standard” language, and connect them with low status or illiteracy. Chambers and Trudgill themselves define dialects as varieties of language, often usefully seen as “subdivisions of a particular language”. The main difference between their use of the term and the “common usage” definitions is that, in Chambers and Trudgill’s definition, all kinds of varieties (including Standard English) may be termed dialects, and by definition everyone speaks at least one dialect. Dialects are distinguished from accents on the basis of the levels of language involved: while “accent” only refers to pronunciation, “dialect” refers to “varieties which are grammatically (and perhaps lexically) as well as phonologically different from other varieties” (Chambers –​Trudgill 1980 [1997: 5]). We might note that geography is not mentioned here at all. Instead, the authors make a difference between dialect geography and urban dialectology, using dialectology as a general cover term for the study of both regional and social variation. Francis (1983), in a book also titled Dialectology, gives a similarly general definition of “dialect”: “varieties of a language used by groups smaller than the total community of speakers of the language” (Francis 1983: 1); however, in

220

Merja Stenroos

practice his discussion is largely focussed on geographical dialects. Chambers and Trudgill, on the other hand, subsume the entire discipline of sociolinguistics –​at that point mainly focussed on the study of urban communities –​under “urban dialectology”, using dialectology as the umbrella term for the study of linguistic variation. In the preface to the 1997 edition, the authors comment on this decision: Our integration of sociolinguistics with more venerable traditions as a highly influential new branch of urban dialectology surprised a few readers but was generally received as an interesting innovation. Now it would be shocking, and hopelessly muddled, if someone tried to keep them apart. (Chambers –​Trudgill 1997: xiii)

From the 1980s onwards, three core senses might be distinguished in the scholarly use of the term “dialect”: a) the general sense equivalent to “variety”, which may include both regiolects and sociolects, and two more specific senses, referring to b) geographical and c) non-​standard varieties respectively. In practice, sense b) normally presupposes c), while the converse is not the case. With the integration of sociolinguistics into dialectology and the distinction of “urban” and “rural” dialects, as well as the widespread use of “dialect” to contrast with “accent”, one might have expected sense a) to take over altogether. Indeed, a brief survey of dictionary resources online shows that several resources, including dictionary.com and wiktionary.org, endorse the more general sense in their main definition: A variety of a language that is characteristic of a particular area, community or social group, differing from other varieties of the same language in relatively minor ways as regards grammar, phonology, and lexicon (wiktionary.org)

When the general sense of “dialect” is referred to, the terms “regional dialect”, “regiolect” and “geolect” are frequently given as subordinate terms. However, it is clear that the simple term “dialect” is still often associated with geographical variation, with several resources listing the geographical sense as the primary one; these include Merriam-​Webster, dictionary.cambridge.org and the online Encyclopædia Britannica: Dialect, a variety of a language that signals where a person comes from. The notion is usually interpreted geographically (regional dialect), but it also has some application in relation to a person’s social background (class dialect) or occupation (occupational dialect) [...] A dialect is chiefly distinguished from other dialects of the same language by features of linguistic structure—​i.e., grammar (specifically morphology and syntax) and vocabulary. (Encyclopædia Britannica)

What, If Anything, Are Middle English Dialects?

221

It may also be noted that Trudgill’s excellent textbook The dialects of England (Trudgill 1999) deals exclusively with geographical variation, while acknowledging the presence of other kinds of dialect, such as ethnic varieties; the point to note here is that, despite its explicit delimitation, the book has not been titled The regiolects or regional dialects of England. The geographical sense is, clearly, felt to be central enough for no specification to be needed. All the present-​day sources here consulted do explicitly agree on the applicability of the term “dialect” to a broad range of uses, even though it is largely associated with geographical variation. In the study of historical periods of English, the geographical aspect is even more dominant. Virtually all references to dialect and dialects in the Wiley-​Blackwell Handbook of the History of English (van Kemenade –​Los, 2009) have to do with geography, even when referring to present-​day data (Tagliamonte 2009: 479–​482). Laing and Lass, in their chapter on Early Middle English dialectology, suggest that “space is only one dimension of dialectology”, noting that spatial distribution involves two other dimensions as well: time and social milieu (“human interactions and the intricacies of language use”) (Laing –​Lass 2009: 417). At the same time, their introduction makes clear that dialectology has to do with geographical distributions, and indeed they define historical dialectology as “historical linguistics with a spatial emphasis”. (Laing –​Lass 2009: 418). Even in the study of present-​day varieties, the term “dialectology” tends to be mostly associated with the geographical sense, despite its broader use by Chambers and Trudgill (1980 [1997]). Perhaps symptomatically, the current Wikipedia entry for “dialectology” defines it as “a sub-​field of sociolinguistics” which deals with variation in language “based primarily on geographic distribution” (wikipedia.org, Dialectology). Here, the relationship between sociolinguistics and dialectology has been reversed compared to the model of Chambers and Trudgill (1980 [1997]), clearly reflecting the growth of sociolinguistics as the mainstream field of research, no longer in the shadow of the “more venerable” traditions. Whichever of the three core senses is involved, the term “dialect” shares a common characteristic with the terms “variety”, “language” and “accent”: it denotes a category that cannot be defined by linguistic means, through the presence or absence of specific linguistic forms. This point is made by Chambers and Trudgill (1980 [1997: 5]): The labels “dialect” and “accent” [...] are used by linguists in an essentially ad hoc manner. This may be rather surprising to many people, since we are used to talking of accents and dialects as if they were well-​defined, separate entities: “a southern accent”, “the Somerset dialect”. Usually, however, this is actually not the case.

222

Merja Stenroos

Chambers and Trudgill explain this fluidity with the continuous nature of geographical variation, with overlapping distributions of individual items: the qualification “usually” implies that, given enough isolation, definable dialects may indeed exist. Kretzschmar (2009: 4) formulates the impossibility of defining labels such as “dialect” linguistically in considerable stronger terms: “[t]‌he notion of the existence of any language or dialect is actually an ‘observational artifact’ that comes from our perceptions of the available variants, at one point in time and for a particular group of speakers”. In other words, dialects and varieties are not real things that we uncover, but labels that we use for convenience to classify what we hear or read. Kretzschmar’s argument goes much further than pointing out the problems in drawing boundaries: here, the whole concept of dialects becomes irrelevant to the study of the actual linguistic variation, being purely a matter of perception. That dialects, regional or otherwise, do not exist as definable entities is not in itself a controversial point, and it is made in most reasonably recent textbooks and handbooks dealing with linguistic variation. However, the implications of this point are not necessarily easy to draw in the study of historical periods of language, such as Middle English, both because of the nature of the surviving evidence and the presence of strong research traditions for which the labelling of dialects is central.

3. Middle English Dialectology 3.1. Geography and Middle English Variation The central role claimed for Middle English dialectology reflects, as noted above, the extreme variability of written Middle English. As Milroy (1992: 156) puts it, “The most striking fact about Middle English is that it exhibits by far the greatest diversity in written language of any period before or since”. This diversity, which involves all levels of language, may be illustrated with a line from the fourteenth-​ century poem Prick of Conscience in six out of the ca 120 surviving manuscripts: (1)

a.  whenne hit ys born hyt sayeth . e . e b. when hit is bore hit seiþ E E c. quenn it is born þen sais it e E d. wan it is born yan says it E e. whan it is ybore it sayþ E . E . f. when hit ys yborne hit seiþ E   ‘when it is born, it says E’

What, If Anything, Are Middle English Dialects?

223

The line refers to newborn baby girls, whose sex may be established by their cry, as they cry out the first letter of the name Eve; newborn boys, as everyone knows, cry “A” for Adam. It may be noted that the spelling of every single word in this line varies between the six manuscripts. Some variation seems to be purely orthographic (i vs y in hit, is, sais vs hyt, ys, says); here, example (a) also shows orthographic variation within a single text (hit vs hyt ‘it’). Other spelling variants might have some correlation to spoken variation (when vs wan, it vs hit). There is also morphological variation (bore vs ybore; sayeth vs sais) and variation in syntax (it sayþ vs þan says it). While this example does not show lexical variation, it of course exists as well, as may be shown by an example from the Cursor Mundi, concerning the question why Adam had such a wonderful name: (2)

Parfay þat ys bot eþe to rede (Bodleian Fairfax 14) For soþe þt is liȝte to rede (Cambridge, Trinity R.3.8) ‘Truly that is easy to understand’

While all other manuscript versions of the Cursor Mundi show northern characteristics, the Trinity version seems to have been translated into a western variety with much lexical substitution: here it has liȝte and for soþe for ‘easy’ and ‘truly’, where the northern versions have eþe and the French form parfay. All this written variation is traditionally referred to as “dialectal”. The thinking has generally been that, in Middle English, written variation was to a very large extent conditioned by geography: At the close of the fourteenth century, the written language was local or regional dialect as a matter of course; typically, the area in which a man acquired his written language can be deduced from the form of the language itself (Benskin 1992: 71) The main source of diversity in written Middle English is regional and local variation (Burrow & Turville-​Petre 1996: 5) [W]‌riters naturally wrote in the dialect of that part of the country to which they belonged (Baugh & Cable 2012: 183)

Geography is also the only variable that has been considered in Middle English dialect surveys, including The Linguistic Atlas of Late Mediaeval English (LALME 1986) and the Linguistic Atlas of Early Middle English (LAEME 2013–​), as well as the earlier surveys by Kristensson (1967–​2001) and Moore, Meech and Whitehall (1935). While the compilers of these surveys acknowledge that other kinds of variation exist, they are not considered dialectal: in the LALME tradition, which here builds on earlier terminology, texts that cannot be placed geographically are considered to contain either “standardised” or “colourless” usage

224

Merja Stenroos

or a Mischsprache or mixed language. For any work related to these surveys, then, the term “dialect” may be assumed to refer to a geographical variety; however, the precise ways in which language and geography are assumed to interact vary greatly and are not always made fully explicit.

3.2. Spoken or Written Dialects? One of the most fundamental differences between approaches to Middle English dialects has to do with the understanding of which medium is being studied: is Middle English dialectology about spoken or written variation, or both? Most scholars assume that the written variation we encounter reflects spoken variation at least to some degree: as Smith (2005: 90) carefully puts it, [W]‌hen ME was employed in the written mode after 1066, it reflected historical changes and dialectal variation which had been disguised by OE written standardisation, with local patterns of spelling (reflecting, albeit conventionally, local pronunciations), grammar and even lexicon.

Beyond this general point, however, scholars vary greatly in their approaches to the written evidence. The mainstream tradition of linguistics through much of the twentieth century considered spoken language the only legitimate goal of linguistic enquiry: writing, in Bloomfield’s (1933: 21) often quoted words, was “not language, but merely a way of recording language”. Accordingly, Middle English written variation would be of interest only in as far as it could be used to extrapolate about the underlying spoken form. Views on the extent to which written Middle English may be considered an accurate record of spoken variation have also varied greatly, from the extremely pessimistic view that written language is “corrupted” by scribes (Labov 1972: 98, Tolkien 1929: 104) to the optimistic view that Middle English spellings may be treated as a “set of field recordings of utterances” (Lass 1997: 65). Most traditional dialectologists have treated spellings as direct evidence for phonology, with the result that it has sometimes been possible to consider the dialect of two texts to be “the same” even when their written form is completely different. The surveys of Moore, Meech and Whitehall (1935) and Kristensson (1967–​) represent this approach, but both refer explicitly to spellings in their discussion of the evidence, and Kristensson lists all spellings systematically, making his findings transparent irrespective of scholarly approach. A completely different approach was suggested by McIntosh (1956, 1963) in one of the most daring and important contributions to the study of historical language. McIntosh noted that much written variation has no necessary connection with spoken variation at all, but that it still exhibits orderly variation: his

What, If Anything, Are Middle English Dialects?

225

contention was that Middle English written variation may be expected to vary in significant ways across geography and may be studied according to the same principles as present-​day spoken language. This principle, which is still controversial in many scholarly contexts, has been widely adopted in the study of historical English variation. Few scholars would, presumably, follow McIntosh in his suggestion that written language could be studied entirely in isolation, but his point that the relationship between written language and its spoken equivalent is comparative rather than descriptive (McIntosh 1956: 12) is now widely accepted. The same point was developed by McIntosh’s fellow LALME compiler, Samuels (1972: 6) into a model that has been adopted in much recent work (cf Bergs 2005: 16; Smith 1996: 17; Black 1999: 63). Samuels sees the written and spoken chains as developing in parallel, neither isolated from nor dependent on each other, but in constant interaction. The implication is that written patterns may tell us much about what is going on in speech, but they do not reflect speech directly.

3.3. Evidence: Documentary and Literary Texts The focus on geographical variation in Middle English gives rise to some major challenges, in addition to those common to all study of early languages. All historical linguists dealing with earlier periods are restricted by the limited amount of data surviving and its likewise limited range, consisting only of written data and deriving from whatever text types survive. In addition to these general limitations, the most immediate challenge in Middle English is the lack of contextual information. Often there is no information about when and where a text was produced. Even when we do have such information, we only exceptionally know anything about the background of the person who wrote or copied the text. The lack of contextual data is much more marked with regard to some materials than others, and it is therefore important to distinguish between two main types of Middle English text: documentary and literary texts. Documentary texts may be defined as texts which have a pragmatic function, such as transferring values, recording decisions or conveying news, and which relate to specific situations and people at a specific point of time. This definition covers a wide range of text types such as wills, contracts, letters, accounts and memoranda, and includes both administrative documents and private texts such as letters. Because of their immediate relevance for a specific historical context, they usually provide a date and often information about provenance:

226 (3)

Merja Stenroos Made at Penreth yo iij day of Septembre ye yhere of kyng Henry sexte eftir ye conquest ixme (Gloucestershire Archives: D326/​T158) ‘made at Penrith the 3 day of September in the ninth year of Henry VI after the Conquest’

As they typically concern a limited number of people, documentary texts are mostly produced in one or two fair copies only; sometimes an extra archive copy is produced in addition, or a draft version is archived. Accordingly, the scribal histories of documentary texts are shallow: for the most part they are products of a single scribe, or copies at a single remove. Documentary texts are usually contrasted with literary texts, here meaning more or less all other texts, not only those usually thought of as “literary” (in the sense of creative, aesthetic or entertaining writing). Literary texts in this broader sense include medical works, cookbooks, treatises, chronicles and sermons as well as romances, allegories and drama. They may be defined as texts that do not relate to a specific situation, but that have a more general application: a medical recipe may be of interest to many different people at different times, just like a poem may. They are typically produced in multiple copies that may be widely dispersed in time and space, and often survive in versions that are the result of several layers of scribal copying. Unlike documentary texts, literary texts for the most part give no clues as to where and when the text was produced or composed; it is generally also unknown how many copyings the text has gone through or what each different scribe has contributed to its language. Literary texts tend to be what most scholars study: apart from the intrinsic value traditionally attached to creative or learned texts, they are much more commonly available in editions. They are also more likely to provide long texts with a good range of vocabulary and grammatical structure, as well as large quantities of data. In contrast, relatively few documentary texts have been edited, and it has only recently become possible to photograph them at archives.1 In addition, many documentary texts are fairly short, and some types of documents (e.g. rentals and accounts) may be rather repetitive in structure and vocabulary (something that may be an advantage as well as a setback, depending on the study). Documentary texts provide much more contextual information than literary texts normally do, making them an important part of virtually any survey of Middle English dialectal variation. However, neither kind of text can provide data such as we collect from living informants: complete with precise 1 A Corpus of Middle English Local Documents (University of Stavanger, www.uis.no/ meld) will make available transcriptions of over 2,000 documentary texts in English from the period 1399–1525. It is currently in process of publication.

What, If Anything, Are Middle English Dialects?

227

and detailed metadata, and sampled according to well-​defined principles. The practices taken for granted in present-​day dialectology are simply not available for Middle English: we always need to reconstruct, fill in gaps and make educated guesses, because we do not have all the facts. Accordingly, our basic assumptions, including the model of linguistic variation on which the analysis is built, play a much more fundamental role than they do in present-​day dialectology.

4. Models of Middle English Geographical Variation 4.1. Some Basic Assumptions: Dialects and Places One of the most fundamental assumptions in the study of Middle English dialects is that variation, written as well as spoken, is on the whole conditioned by geography. Geography has generally formed the only large-​scale framework against which variation has been studied, and texts have been excluded from the surveys on the grounds of being either geographically mixed or geographically neutral (referred to variously as “supralocal”, “colourless” or “standardised”). One of the traditional assumptions about dialects both among scholars and lay people is that every place or area has its “own” dialect. Both in Middle English and in the present day such dialects are often perceived as being endangered or disappearing. The assumption that a dialect should exist –​or have existed –​in a “real” or “pure” form, of course, presumes exactly the kind of idea of dialects as entities that scholars now, at least in principle, agree is wrong; it is, however, deeply ingrained in the study of Middle English variation. Even when it is accepted that “pure” dialects do not exist, the idea that particular combinations of linguistic forms belong to specific places or areas is fundamental to most work on Middle English; the problems relating to this idea will be discussed further in section 6 below. In the study of Middle English dialectal variation, three major models may be distinguished: the dialect area, the dialect continuum and what we might call the social space model. All of these models are also found in the study of present-​ day geographical variation. However, when studying variation in the present, we do not need to reconstruct, and can therefore simply use these models to describe geographical patterns that we already know. In Middle English, on the other hand, these models tend to become the actual means that we use to organize the data –​rather than describing patterns, we at least to some extent create them using these models.

228

Merja Stenroos

4.2. The Dialect Area Model Descriptions of Middle English based on the dialect area model are found in virtually any textbook that includes a chapter on Middle English. The geographical area where English was spoken is divided into distinct areas on the map, each one of them representing the approximate area where a specific dialect of the language is or was spoken. The dialect itself is defined with a list of distinctive forms, or a diagnostic feature set (Alcorn et al. 2019: 3). There is no necessary expectation that all speakers in the area will use all the forms, or that no speakers elsewhere will; however, their combination is seen as typical of and/​or defining for the particular dialect area in question. For example, Smith (2005: 119) lists the following features as typical of the Northern dialect(s) of Middle English: • a spellings for OE ā, e.g. stane stone • ui, uCe spellings for OE ō, e.g. guid/​gude good • -​ es/​-​is type inflexions for both 3rd person present singular and present plural of verbs, e.g. standis stands • -​ and as the present participle inflexion • scho for PDE she • early occurrence of th-​type 3rd personal plural pronouns in all cases, e.g. thay, thaim, thair • early loss of weak/​strong adjective distinction, e.g. the strang strif ‘the violent struggle’, the strif is strang • distinctive northern vocabulary, e.g. til to, fra from • y and þ written identically, as y The most commonly reproduced dialect area map was originally provided in the classic textbook by Baugh and Cable, first published (with Baugh as the author) in 1935. This map is divided into the Northern, West Midland, East Midland, Southern and Kentish dialect areas, and was based on the then newly completed survey by Moore, Meech and Whitehall, commissioned by the Middle English Dictionary project. It may be noted that the Old and Middle English dialect maps provided by Baugh and Cable make use of quite different dialect area labels and divisions, causing generations of students (and perhaps their instructors too) confusion and worry as they have tried to explain how and why the Mercian dialect disappeared and was replaced by the West and East Midland dialects. The perceived change, of course, merely reflects the fact that the two maps are based on different principles. The Old English dialect areas (Northumbrian, Mercian,

What, If Anything, Are Middle English Dialects?

229

West Saxon and Kentish) are based on the divisions suggested by Sweet (1876, see Hogg 2009: 396–​397 for a discussion), who defined them on the basis of what was known or assumed about the tribal and political boundaries of Anglo-​Saxon kingdoms. While the equation of linguistic boundaries with tribal ones has been criticized by later scholars, it was probably the most sensible option given the limited evidence: the problem, rather, lies in the reification of the dialects into stable entities that suddenly, somehow, change into different ones in the Middle English period. The Middle English boundaries came to be different not only because the political map had changed, but because they were based on actual survey data. The results of the survey by Moore, Meech and Whitehall were published in 1935 as “Middle English dialect characteristics and dialect boundaries”. The survey was based on 44 literary texts that could be localized on external grounds and 231 explicitly localized documentary texts. It should be noted that they only included texts containing language which corresponded to what they expected of each dialect; they also generally avoided texts that were localized close to what they assumed to be the dialect boundaries. In the dialect area model, the main aim of the dialectologist is to determine the dialect boundaries. The best kind of boundaries are those which follow bundles of isoglosses: that is, coinciding distribution boundaries of several dialectal forms. However, as such bundles are not very common in the English language area, most boundaries simply follow the distributions of features that the dialectologist considers salient. Moore, Meech and Whitehall (1935) defined eleven features, eight of which provided coherent enough geographical distribution to enable them to draw “isophonic lines” dividing the country into ten areas. Moore, Meech and Whitehall were criticized by both McIntosh and Strang for using a too small set of features to define dialect areas. The criticism might have been somewhat unfair considering the aims of the survey: Moore, Meech and Whitehall set out to define dialect areas for a practical, diagnostic purposes, not to provide a large-​scale survey of dialectal variation such as was the aim of LALME. It has been pointed out, for example by Upton (2006: 311), that dialect areas can in fact only be established on the basis of relatively few criteria, as larger numbers of criteria unavoidably clutter the picture. The question is, then, to what extent dialect areas make sense at all. Referring to general areas such as “the north” and “the west” is universal practice in the study of Middle English, and useful as long as the dialects are not interpreted as absolute entities. Whenever we find, for example, a Middle English text that contains a set of typical northern features, such as those listed above, with nothing directly contradicting them, we say that it is written in a northern

230

Merja Stenroos

dialect; this classification may then lead us to assumptions about the provenance of that particular text. It is when such basic diagnostics are interpreted as “scientific”, turning both dialects areas and the localizations of texts into empirical facts, that problems arise.

4.3. The “Dialect Continuum” Model Categorizing dialects using the area model, based on a few salient markers, is the intuitive procedure of scholars and laypeople alike when faced with a sample of unfamiliar speech or (variable) writing. The well-​known fictional character Henry Higgins in Shaw’s Pygmalion, however, claimed considerably more refined skills of localization: “You can spot an Irishman or a Yorkshireman by his brogue. I can place any man within six miles. I can place him within two miles in London. Sometimes within two streets”. This citation was used by Benskin (1991: 9) to introduce a discussion of the principles of localization developed by McIntosh and used in the compilation of LALME. The “fit-​technique” developed by McIntosh built on a very similar procedure to the one presumably used by Professor Higgins: comparing texts in minute detail in order to produce a typology where more texts could be fitted in.2 The LALME methodology turned the geographical localization of texts into the main research question. Texts, including a very large number of literary ones, were localized in relation to each other on the basis of their linguistic forms, using documentary texts as anchors to relate the localizations to the geographical map. The result was an Atlas with more than a thousand survey points consisting of texts (mostly) dated to the period 1350–​1450; the same technique was later used to produce the Linguistic Atlas of Early Middle English (LAEME; Laing 2013–​). The fit-​technique is based on the model of the dialect continuum. The term refers to the well-​known fact that geographically adjacent varieties of (usually historically related) language tend to be similar to each other, and that the similarity tends to decrease with distance (see Chambers –​Trudgill 1980 [1997: 5–​7] for a discussion). The main assumption behind the fit-​technique is that this continuum is a general characteristic of linguistic variation across space and that it works in a more or less regular, uninterrupted way, making it possible to localize individual informants whose origins are unknown. In other words, the continuum is made into an analytical rather than descriptive tool.

2 McIntosh himself claimed an accuracy to within ten miles of “the correct absolute position” of each text in the areas with best “coverage” (LALME I: 12).

What, If Anything, Are Middle English Dialects?

231

Applying this tool to the Late Middle English materials presupposes further assumptions. The most important one is that linguistic variation in written Middle English also forms a regular dialect continuum. This implies that written Middle English varies in the same way that spoken dialects (ideally) do, since the idea of a dialect continuum is based on spoken variation. It also implies that the written language of any given place is reasonably homogeneous, so that every informant that represents a local dialect may be expected to fit into the continuum: accordingly, a text that does not fit in will either be a Mischsprache or represent another continuum (such as Hiberno-​English) (LALME I: 12). The requirement of “fit” is given as an argument for the reliability of the method: if a text can be placed in relation to other, already “localized” ones, this assures its genuinely local character. The other side of this argument is that, once the overall framework of localizations is established, those texts which do not fit in are discarded. It also means that the anchor texts –​the documents with known provenance that are used to connect the continuum to the physical map –​have to be selected so that they fit in with the overall continuum, just like Moore, Meech and Whitehall only included those texts which answered to their general expectations. The resulting network of localizations thus forms an unbroken continuum in which the dialectal forms of each location, in the absence of actual informants, may be reconstructed (see e.g. Williamson 2000, 2004). The continuum is based purely on linguistic similarities: the actual historical context of texts is taken into account only in as far as it agrees with the established linguistic patterns, and a large part of the available material is discarded as dialectal evidence. The main difference between the LALME method and any present-​day dialect survey is, accordingly, that the former to a large extent relies on localizations made on the basis of the language itself, prioritizing linguistic fit over known provenance. Here it will be worth looking at two examples of texts for which “dialect” and provenance seem to differ. The first example is a sample from a 1472 account from Wimborne Minster in Dorset (Dorchester, Dorset History Centre: PE/​WM/​CW/​1/​32), listing various expenses in connection with maintenance and renovation work on the minster: (4)

In expences & costis In primis pro j glassere wene he come [...] Itm’ for [...] qwen the lente clothe was heget vppe ~ ij d [...] Itm’ payt to harr’ hunt for atte the newe wyndowe be v days [...] ij s vj d Itm’ payt to the same harr’ for hongging’ of the trhede bell ~ ij s’ viij d Itm payt for the glassier’ys tabull and his by v days ~ xx d Itm’ payt for hor horsemete ~ ij s iiij d [...] Itm’ for wodde ~ ij d [...]

232

Merja Stenroos Itm’ payt to j mon be ij tymys aythere tymys ij days for diuers nessessare [...] Itm’ payt for ij torchis eboghte pryce x s iiij d ‘In expenses and costs In primis pro one glazier during his visit[...] Item, for [...] when the lent cloth was hung up, 2 pence [...] Item, paid to Harry Hunt for [work] on the new window for 5 days, 2 shillings and sixpence, Item, paid to the same Harry for hanging up the third bell, 2 shillings and 8 pence, Item, paid for the glazier’s table, and that of [Harry Hunt’s], for five days, 20 pence, Item, paid for feeding their horses, 2 shillings and 4 pence [...] Item, for wood, 2 pence [...] Item, paid to one man at two occasions, each time for two days, for diverse necessary work [...] Item, paid for two torches bought, price 10 shillings and 4 pence’

Map 1:  Provenance (diamond) and approximate eLALME “fit” (circle) of the Wimborne account

The text, which is not listed in LALME, contains three dialectal forms that do not seem to suggest Dorset at all, but rather indicate an area much further north: qwen when, hor their and mon man. A search for the combination of

What, If Anything, Are Middle English Dialects?

233

these forms in eLALME suggests a limited area including east Cheshire and northern Staffordshire and Derbyshire (see Map 1). Most other linguistic forms found in the text agree both with this localization and with the provenance in Dorset. While the form payt paid was not collected in LALME, the MELD corpus shows this form mainly in documents from the northern and northwestern areas, including Cheshire; however, it also appears in two other church accounts from the south (Dorset and Sussex). A single form, eboghte in a payment for ij torchis eboghte ‘for two torches bought’ suggests a remnant of the Old English prefix ge-​, generally limited to the southern dialects in late Middle English. The LALME principle is to assume the simplest possible dialectal makeup (cf Benskin & Laing 1981: 84): the most logical procedure here would then be to localize the text in the northwestern area and consider the single clearly “southern” form, eboghte, a relict form (see LALME I: 13–​14). Another example is the 1473 Ordinance of the Guild of All Saints in Cambridge (Oxford, Bodleian Rawlinson C.541), discussed in detail by Bergstrøm (2017: 191–​193; see also Bergstrøm 2013, passim). In his study of the late and post-​medieval dialect materials of Cambridge, Bergstrøm studied 169 documentary texts, all of which were (as far as may be established) actually produced in Cambridge. In this corpus, the All Saints text is the only one which stands out with regard to its linguistic forms, as shown in the following sample: (5)

Also it ís ordeynyde that qwhan the maystyrs schall’ receyue the catell’ of the gylde into thar handys yche of tham schall’ fynde . ij . suffycient plegges bowndyn wyth thame in a syngyll oblygacione for to make a trew delyuerance ageyne off syche goodys as thay receyve wyth the increce cumynge therof at the nexte generall’ day folowynge beforne ye alderman . Also the alderman /​schall’ haue at every generall’day to hys drynke and for hys gestys a galone of aale and euery maystyr a pottell~ the clarke a+potell’ and the deen . a qwarte (fols 6v–​7r) ‘Also it is ordained that when the Masters receive the guild’s funds into their hands, each of them must provide two sufficient pledges for which they are bound in a single obligation to deliver faithfully the funds that they received, as well as the increase gained, at the following general day before the Alderman. Also, the Alderman shall have at every general day a gallon of ale for himself and his guests, and every Master a pottle, the Clerk a pottle and the Dean a quart’.

The text contains a number of features that would agree with a somewhat more northern provenance, including a full set of th-​forms in the third person plural pronoun (thay, thar, tham(e)) and regular q-​spellings for wh-​ (qwhan when); the spelling ye the, showing merger of and realized as , is at this point of time also still found mainly to the north of Cambridge. The text was listed as an “associated manuscript” under Cambridgeshire in the LALME index,

234

Merja Stenroos

with the comment “Language not of Cambridge, but more like that of S Lincs, Soke or Rutland”. In his study, Bergstrøm (2017: 193) localized it, using eLALME, on the basis of four linguistic forms (qw-​for wh-​, noth not, thar their and theis these), in a small area around the boundary between South Lincolnshire and North Norfolk (see Map 2).

Map 2:  Provenance (diamond) and approximate eLALME “fit” (triangle) of the Ordinance of the Guild of All Saints (after Bergstrøm 2017: 193)

With regard to these two texts, we might pose the question referred to at the outset: is it the dialect or the geography that matters? The “dialect” here refers to the fit into the LALME continuum, while the geography refers to the actual provenance of the texts. The question is, of course, unanswerable: what matters clearly depends on the purpose. What does it mean that a text is localized in a particular place using the fit-​ technique? As the localization is based on a linguistic comparison, it basically tells us with which other materials the text groups linguistically: where it fits in relation to other texts in a perfect dialect continuum. We may say that the Wimborne account seems to be (mainly) an example of a Northwest Midland

What, If Anything, Are Middle English Dialects?

235

dialect, and that the All Saints text was probably produced by someone with linguistic habits agreeing with the Lincolnshire-​Norfolk border area. From this point of view, the fact that the texts were beyond doubt produced in Dorset and Cambridge respectively, is not relevant: placed on a LALME map, they would form evidence for the northwestern and Lincolnshire/​Norfolk areas, not for their places of provenance. The LALME approach clearly presupposes that the localizations reflect actual geographical connections at some level; at the same time, as the localizations are strictly relative, they cannot be used as direct evidence for any such real-​ life connections. In contrast, known historical provenances provide evidence of what was in fact going on linguistically in particular places: the texts here considered, for example, are proof that certain linguistic forms were produced in Dorset and Cambridge respectively, and were considered acceptable in an official text used by local people, irrespective of whether or not they were produced by a scribe with a local background. These geographical connections are very different from those suggested by the fit-​technique, and certainly answer different research questions; in particular, the fact that they relate to historical contexts makes them more useful for a sociolinguistically oriented study.

4.4. The Social Space Model The problem with both models discussed above is that they are idealizations: we do not normally find boundaries between distinct dialect areas and spatial variation generally does not form a regular continuum in all directions. As David Britain (2002: 604) has pointed out, linguistic geography does not have to do with physical distance as such, but with spatiality, the interaction of this distance (“Euclidean space”) with both perceived space and social space: “the space shaped by social organization and human agency, by the human manipulation of the landscape, by the contextualization of face-​to-​face interaction” (Britain 2002: 604). People do not interact with each other in amounts directly proportional to physical distance, either in speech or writing: rather, their interaction depends on networks to do with commerce, education, religion, employment and so on, as well as on available travel routes. We might therefore expect linguistic similarities to reflect contact rather than distance in itself: affinities in social space rather than in the strictly Euclidean space depicted on a map. The concept of social space explains why linguistic forms on a present-​day map may cluster and suggest changes that leap over space, as in the phenomenon known as city-​hopping, something that Thengs (2013: 298) has shown also takes place in

236

Merja Stenroos

Middle English documentary texts. As Kretzschmar (2009) has pointed out, it is this kind of clustering and leaping, rather than the regular patterns implied by the dialect area and dialect continuum models, that we would expect of complex systems such as language. We cannot, accordingly, expect to reconstruct “real” historical dialect patterns by organizing the linguistic data into regular areas or continua. Instead, we have to expect a more complex picture, reflecting the actual communities and networks within or through which language changes are disseminated: in other words, bringing together the spatial and social dimensions. For that purpose, we need to take as a starting point whatever extralinguistic information we have.

5. The Importance of Provenance: Who Wrote What Where Unlike literary texts, documentary texts may for the most part be related to actual locations, whether through explicit localizing clauses or through inferences based on names or manuscript associations. Accordingly, they allow us to study the language actually produced in a given place. There can, of course, be no certainty that a text from a particular place was written by someone native to, or permanently resident in, that place. It may be –​and has been –​argued that scribes and clerks travelled, and it is not difficult to find texts with linguistic forms unexpected for their provenance, as in the two examples discussed in section 4.3 above. However, there are at least two reasons why this does not diminish the importance of historical provenance. First of all, most texts produced locally (unlike, for example, texts produced by central government offices) are likely to reflect local usage. In the Cambridge material studied by Bergstrøm (2017), only one text out of 169 stood out as being linguistically substantially different. This homogeneity is not a product of the corpus compilation: the texts were collected from archives with no attention to dialect, and everything that was found in English from the survey period was included.3 There is no reason to doubt that most scribes working in a given place would have been local or from nearby, even though larger cities certainly received much migration from further afield. The last point leads to the second reason. Migration, mobility and contact play a crucial role in the dynamics of linguistic variation and change and would certainly have done so in the late medieval period. To include texts which show their effects, rather than just those which seem to represent “unmixed” dialects, 3 It may be noted that only two of the 169 texts were also used in LALME, reflecting the low proportion of documentary texts included in the southern part of the LALME survey.

What, If Anything, Are Middle English Dialects?

237

is, it might be argued, crucial for making sense of the linguistic developments, both spoken and written. For example, the Ordinance of the All Saints guild in Cambridge, despite its untypical dialect features, was an important text for its community: it had considerable authority and was continued and added to over decades. Not to include it in a study of “the Cambridge dialect” means disregarding an influential part of the actual textual output of Cambridge, of potential importance for the area’s linguistic and social development. The compilation of A Corpus of Middle English Local Documents (MELD) at the University of Stavanger is based on the principle that every text with an extralinguistic connection to a place, irrespective of its language, is equally valid as evidence; the localizations here are based purely on provenance. Considering the arguments above, it might be of interest to ask to what extent such a corpus might show coherent patterns at all: even assuming that most texts were produced by local scribes, one might expect a dialect map which includes all the noise excluded from LALME to show very diluted geographical patterns, if any at all.

Map 3:  The distribution of hor, hore THEIR in MELD

238

Merja Stenroos

This does not, however, turn out to be the case. Map 3 shows the distribution of all occurrences of hor, hore their in MELD. With one single exception, all the texts that contain this form are from a clearly defined area in the Northwest Midlands (including the Isle of Man), with a particularly dense concentration in the Cheshire–​South Lancashire area. The distribution is, in fact, considerably more coherent than that of the same forms in eLALME. The eLALME map generated for the forms hor and hore shows the same main area of distribution in the northwest Midlands, including Cheshire, Derbyshire, South Lancashire and Staffordshire. In addition, however, there is a large band of hor(e) forms spread across the south of England, from Herefordshire and Gloucestershire in the west to Essex and London in the east. Some of these more southern texts are earlier than the MELD texts and may represent an earlier distribution; however, many are from the fifteenth century. As most LALME texts are literary texts which may have been copied from other dialects and show relict usage, a large number of “outliers” is not surprising: in many cases, they probably represent minority usage, although this is difficult to ascertain from the LALME material. Map 3, on the other hand, shows only one outlier: the Dorset text discussed in 4.3 above. As with the Cambridge material, there seems to be very little noise, despite all potentially mobile scribes. Apart from the greater coherence of the pattern in MELD, the two maps are very similar. Other comparisons between maps based on MELD and LALME localizations have shown similar results: so hit it and gud good in Thengs (2013: 102–​105) and mon man in Stenroos & Thengs (2020: 90). Some features, however, show much greater differences between the MELD and LALME distributions: this is particularly the case with purely orthographic features such as the spellings corresponding to present-​day and (see Thengs 2013: 117–​126; cf also Stenroos 2004: 276–​279, Jensen 2010: 374–​75 and 2012; Stenroos –​Thengs 2012). Such major differences reflect, above all, the important fact that not nearly all variation in written Middle English can be sensibly related to geography: the orthography of literary texts, which dominate in LALME, shows highly significant differences from that of documentary texts, probably reflecting the largely (if not entirely) distinct scribal communities enganged in the production of these two types of text (see Stenroos, forthcoming for a discussion). Even when the patterns are encouragingly similar, however, the different principles of localization need to be taken into account. This is, in fact, the case even within the MELD corpus: even where localizations are seemingly solid, the idea of geographical –​or spatial, or regional –​variation is far from straightforward when early historical materials are concerned. This leads to the final point of this

What, If Anything, Are Middle English Dialects?

239

chapter: that a dialect map can never give a definitive picture of “Middle English dialects”. Just like physical geography is not the only independent variable in written Middle English –​or for that matter in any spoken or written language –​it is not a single variable in itself. There is no quintessential way in which linguistic forms relate to geography: the places themselves don’t speak or write. Any speaker connects to geography in many ways, and the same is true of texts and writers. In a present-​day dialect survey, the connection is usually kept constant, and is typically defined as the place where an informant grew up and still lives. For historical surveys, different kinds of information may be available for different texts. Within the MELD corpus, three main kinds of connection between the texts and the geography are distinguished. Some texts provide explicit localizations, such as “given at Penrith” (see ­example 3). Others have a solid historical context: the account of Wimborn Minster (­example 4) was produced in Wimborn, Dorset, and the Beverley Town Cartulary was produced at Beverley. Still others may be inferred (with more or less certainty) from the people and places involved in the document. Sometimes, the same text may be placed in different locations depending on the criteria we use: for example, a Yorkshire nobleman may produce his will in southern England, prior to a dangerous journey to France. By changing the criteria, it is possible to produce a large number of different dialect maps, depending on the available information and the research questions. Each of them shows one particular geographical constellation; however, none of them is the definitive dialect map, simply because there is no such thing as the definitive dialect, or even the definitive localization.

6. Concluding Remarks The aim of this chapter has not been to suggest that the dialect area and dialect continuum should not be used as models. Both are highly useful for working out likely provenances of unlocalized texts; in addition, the LALME method is a fantastic tool for describing how different texts relate to each other linguistically. What is important, however, is to be clear that these models do not –​and are not meant to –​tell us what was going on in the real landscape: who was writing what to whom in which place. The problem is not that the Middle English written variation is viewed from different angles or with different underlying assumptions: problems arise when the constellations assume an absolute existence of their own. This means that being aware of the criteria behind a dialect map is crucial. A LALME map that has placed a text in Worcestershire using the fit-​technique certainly tells

240

Merja Stenroos

us something important and interesting about the language of that text. It does not, however, identify a historical scribe who went to school or worked in Worcestershire. Conversely, when we study a set of documents that were produced in Cambridge, we are not claiming that they were all produced by native Cambridge people or that they all represent something that would be perceived as “the Cambridge dialect” either by their medieval readers or by a modern dialectologist. Despite of this, it may be argued that there is a very real reason to care about provenance, and about the specific connections on which it is based. An approach to linguistic variation that is based on real historical contexts, and that does not worry about pure and mixed dialects, might allow us to see patterns that are unexpected and that might go unnoticed if we focus only on “good dialect evidence”. Such unexpected findings might involve the presence of people from further north in Cambridge or Dorset, or indeed in London; they might also show the literary texts in a different light. This also means that there are no useless texts, just like there are no useless living informants.

References Alcorn, Rhona –​Joanna Kopaczyk –​Bettelou Los –​Benjamin Molineaux. 2019. “Historical dialectology and the Angus McIntosh legacy”. In: Rhona Alcorn –​ Joanna Kopaczyk –​Bettelou Los –​Benjamin Molineaux (eds.): 1–​16. Rhona Alcorn –​Joanna Kopaczyk –​Bettelou Los –​Benjamin Molineaux (eds.). 2019. Historical dialectology in the digital age. Edinburgh: Edinburgh University Press. Baugh, Albert C. –​Thomas Cable. 2012. A history of the English language. 6th edn. Abingdon: Routledge. Baugh, Albert C. & Thomas Cable. 2012. A history of the English language. Benskin, Michael. 1991. “The “fit”-​ technique explained”. In: Riddy, Felicity (ed.): 9–​26. Benskin, Michael. 1992. “Some new perspectives on the origins of standard written English”. In: van Leuvensteijn, Jan A. –​J.B. Berns (eds): 71–​105. Benskin, Michael –​Margaret Laing. 1981. Translations and Mischsprachen in Middle English manuscripts. In: Benskin, Michael–​Michael L. Samuels (eds.): 55–​106. Benskin, Michael–​Michael L. Samuels (eds.). 1981. So meny peple longages and tonges: a Festschrift to Angus McIntosh. Edinburgh: Middle English Dialect Project.

What, If Anything, Are Middle English Dialects?

241

Bergs, Alexander. 2005. Social networks and historical sociolinguistics: studies in morphosyntactic variation in the Paston letters (1421–​1503). (Topics in English linguistics 51). Berlin: Mouton de Gruyter. Bergstrøm, Geir. 2013. Two Cambridge guild ordinances based on the same template: an edition of the Ordinances of the Guilds of St. Clement and All Saints. MA thesis, University of Stavanger. Bergstrøm, Geir. 2017. “Yeuen at Cavmbrigg”: a study of the medieval English documents of Cambridge. PhD thesis, University of Stavanger. Black, Merja. 1999. Parallel lines through time: speech, writing and the confusing case of she. Leeds Studies of English 30: 59–​81. Blake, Norman (ed.). The Cambridge History of the English Language. Vol. II. 1066–​1476. Cambridge: Cambridge University Press. Bloomfield, Leonard, 1933. Language. New York: Holt, Rinehart and Winston. Britain, David. 2002. Space and spatial diffusion. In: Chambers, J. K. –​Peter Trudgill –​Nathalie Schilling-​Estes (eds.): 603–​637. Burrow, John A. –​Thorlac Turville-​Petre. 20053. A book of Middle English. Oxford: Blackwell. Chambers, J. K. & Peter Trudgill. 1997 [1980]. Dialectology. Cambridge: Cambridge University Press. Chambers, J. K. –​Peter Trudgill –​Nathalie Schilling-​Estes (eds.). 2002. Handbook of language variation and change. Oxford: Blackwell. Dossena, Marina –​Roger Lass (eds.). 2004. Methods and data in English historical dialectology. Bern: Peter Lang. Edlund, Lars-​Erik –​Gunnar Persson (eds). 1990. Language: the time-​machine. Stockholm: Almqvist & Wiksell eLALME = Benskin, Michael –​Margaret Laing –​Vasilis Karaiskos –​Keith Williamson (eds.). 2013–​. An electronic version of a Linguistic Atlas of Late Mediaeval English. http://​www.lel.ed.ac.uk/​ihd/​elalme/​elalme.html [accessed 20 March 2021]. Francis, W. N. 1983. Dialectology: an introduction. London: Longman. Hogg, Richard. 2009. “Old English dialectology”. In: Kemenade, Ans van –​ Bettelou Los (eds): 395–​416. Jensen, Vibeke. 2010. Studies in the dialect materials of medieval Yorkshire. PhD thesis, University of Stavanger. Jensen, Vibeke. 2012. “The consonantal element (th) in some late Middle English Yorkshire texts”. In: Tyrkkö, Jukka –​Matti Kilpiö –​Terttu Nevalainen –​ Matti Rissanen (eds.): https://​varieng.helsinki.fi/​series/​volumes/​10/​jensen/​ [accessed 20 March 2021].

242

Merja Stenroos

Kemenade, Ans van –​Bettelou Los (eds). 2009. The handbook of the history of English. Chichester: Wiley-​Blackwell. Kemenade, Ans van –​Bettelou Los (eds). 2009. The handbook of the history of English. Chichester: Wiley-​Blackwell Kretzschmar, William A. Jr. 2009. The linguistics of speech. Cambridge: Cambridge University Press. Kristensson, Gillis. 1967. A survey of Middle English dialects 1290–​1350: the six northern counties. Lund: University of Lund Press. Kristensson, Gillis. 1987. A survey of Middle English dialects 1290–​1350: the West Midland counties. Lund: University of Lund Press. Kristensson, Gillis. 1990. “Middle English kirk and cognates in the Midlands”. In: Edlund, Lars-​Erik –​Gunnar Persson (eds): 51–​61. Kristensson, Gillis. 1995. A survey of Middle English dialects 1290–​1350: the East Midland counties. Lund: University of Lund Press. Kristensson, Gillis. 2001a. A survey of Middle English dialects 1290–​1350: the southern counties. I: vowels (except diphthongs). Lund: University of Lund Press. Kristensson, Gillis. 2001b. A survey of Middle English dialects 1290–​ 1350: the southern counties. II: diphthongs and consonants. Lund: University of Lund Press. Labov, William. 1972. Sociolinguistic patterns. Philadelphia: University of Pennsylvania Press. LAEME = Laing, Margaret. 2013–​. A Linguistic Atlas of Early Middle English.Version 3.2. University of Edinburgh. http://​www.lel.ed.ac.uk/​ihd/​ laeme2/​laeme2.html [accessed 20 March 2021]. Laing, Margaret and Roger Lass. 2009. Early Middle English dialectology: problems and prospects. In Kemenade, Ans van –​Bettelou Los (eds.): 417–​451. LALME = McIntosh, Angus –​Michael L. Samuels –​Michael Benskin. 1986. A Linguistic Atlas of Late Mediaeval English. 4 vols. Aberdeen: Aberdeen University Press. Lass, Roger. 1997. Historical linguistics and language change. Cambridge: Cambridge University Press. Leuvensteijn, Jan A. van –​J.B. Berns (eds). 1992. Dialect and Standard Language in the English, Dutch, German and Norwegian Language Areas. Amsterdam: Koninklijke Nederlandse Akademie van Wetenschappen. McIntosh, Angus. 1956. “The analysis of written Middle English”. Transactions of the Philological Society 55(1): 26–​55.

What, If Anything, Are Middle English Dialects?

243

McIntosh, Angus. 1963. “A new approach to Middle English dialectology”. English Studies 44: 1–​6. MELD = Stenroos, Merja –​Kjetil V. Thengs –​Geir Bergstrøm (eds.). 2017. A Corpus of Middle English Local Documents. Version 2017.1. University of Stavanger. www.uis.no/​meld [accessed 20 March 2021]. Milroy, James. 1992. “Middle English dialectology”. In: Blake, Norman (ed.): 156–​206. Moore, Samuel –​Sanford B. Meech –​Harold Whitehall. 1935. “Middle English dialect characteristics and dialect boundaries”. Essays and studies in English and comparative literature: 1–​60. Mugglestone, Lynda (ed.), The Oxford history of English. Oxford: Oxford University Press. OED = Oxford English Dictionary Online. Oxford: Oxford University Press. http://​www.oed.com [accessed 20 March 2021]. Riddy, Felicity (ed.). 1991. Regionalism in late medieval manuscripts and texts. Cambridge: D.S. Brewer. Samuels, Michael L. 1972. Linguistic Evolution. Cambridge: Cambridge University Press. Smith, Jeremy J. 1996. An historical study of English: function, form and change. London: Routledge. Smith, Jeremy J. 20052 [1999]. Essentials of early English: an introduction to Old, Middle and Early Modern English. London: Routledge. Stenroos, Merja. 2004. “Regional dialects and spelling conventions in Late Middle English: searches for (th) in the LALME data”. In: Dossena, Marina –​ Roger Lass (eds.): 257–​285. Stenroos, Merja. 2020. “Regional variation and supralocalization in late medieval English: comparing administrative and literary texts”. In: Stenroos, Merja –​ Kjetil V. Thengs (eds.): 95–​128. Stenroos, Merja –​Kjetil V. Thengs. 2012. “Two Staffordshires: real and linguistic space in the study of late Middle English dialects”. In: Tyrkkö, Jukka –​Matti Kilpiö –​Terttu Nevalainen –​Matti Rissanen (eds.): http://​www.helsinki.fi/​ varieng/​journal/​volumes/​10/​stenroos_​thengs/​ [accessed 20 March 2021]. Stenroos, Merja –​Kjetil V. Thengs. 2020. “The geography of Middle English documentary texts. In: Stenroos, Merja–​Kjetil V. Thengs (eds): 69–​92. Stenroos, Merja–​Kjetil V. Thengs (eds). 2020. Records of real people: linguistic variation in Middle English local documents. Amsterdam: John Benjamins. Strang, Barbara. 1970. A history of English. London: Methuen.

244

Merja Stenroos

Sweet, Henry. 1876. “Dialects and prehistoric forms of Old English”. Transcactions of the Philological Society 16: 543–​69. Taavitsainen, Irma –​Terttu Nevalainen –​Päivi Pahta –​Matti Rissanen (eds.). 2000. Placing Middle English in context. Berlin: Mouton. Tagliamonte, Sali A. 2009. “Historical change in synchronic perspective: the legacy of British dialects”. In: van Kemenade, Ans –​Bettelou Los (eds): 477–​506. Thengs, Kjetil V. 2013. English medieval documents of the North-​ West Midlands: a study in the language of a real space text corpus. PhD thesis, University of Stavanger. Tolkien, J.R.R. 1929. Ancrene Wisse and Hali Meiðhad, Essays and Studies 14: 104–​126. Trudgill, Peter. 1999. The dialects of England. Oxford: Blackwell. Tyrkkö, Jukka –​Matti Kilpiö –​Terttu Nevalainen –​Matti Rissanen (eds.). 2012. Outposts of historical corpus linguistics: from the Helsinki Corpus to a proliferation of resources. Helsinki: VARIENG. Upton, Clive. 2006. Modern regional English in the British Isles. In Mugglestone, Lynda (ed.): 305–​333. Williamson, Keith. 2000. “Changing spaces: linguistic relationships and the dialect continuum”. In: Taavitsainen, Irma –​Terttu Nevalainen –​Päivi Pahta –​ Matti Rissanen (eds.): 141–​179. Berlin: Mouton. Williamson, Keith. 2004. “On chronicity and space(s) in historical dialectology”. In: Dossena, Marina –​Roger Lass (eds.): 97–​136.

Brita Wårvik

The Demise of Ambiguous Adverb/​ Conjunctions and Manuscript Variation: A Case Study of Tho, Then and When in the Chronicle of Robert of Gloucester In memory of Matti Rissanen (1937–​2018)1

Abstract: One interesting case of the general trend in the development of English adverbial connectives towards less polyfunctionality is the replacement of the Old English ambiguous adverb/​conjunctions þa and þonne by the adverb then and the conjunction when during the Middle English period (1100–​1500). Though the rise of when as a subordinator has been studied quite a lot, the details of the changes leading to the redistribution of the syntactic and discourse-​pragmatic functions of the Old English items are still unclear. The aim of this study is to shed light on the changes by investigating variation in the use of the items in different manuscript versions of one popular Middle English text, known as the Chronicle of Robert of Gloucester. The findings show that one of the manuscripts, Oxford, Bodleian Library, Digby 205, is more advanced in the use of then and when in all discourse contexts as well as in the occurrence of other progressive features, which tallies with earlier studies identifying that version as different also in other respects. To answer the intriguing question to what extent the various changes in Middle English are parts of some major typological change in the language would require many more studies of individual texts, but this is a small contribution towards that goal. Keywords: temporal connectives, Middle English, decrease of polyfunctionality, discourse-​ pragmatic functions, syntactic functions, Chronicle of Robert of Gloucester

1. Introduction Summarising the findings of studies of adverbial subordinators in 49 European languages, Kortmann notes “a clear tendency of a decrease in polyfunctionality over time, be it semantic or syntactic” (1998: 216; original emphasis), which is

1 This article started as part of the project on English adverbial connectives led by the late professor Matti Rissanen. I wish to express my gratitude to him and the project members for inspiring discussions. I also wish to thank the ICOME11 participants and the two anonymous referees for their feedback. All the remaining shortcomings are obviously mine.

246

Brita Wårvik

amply evidenced in the drastic changes in the repertoire of adverbial connectives in English (Rissanen 2012). Among the many changes taking place during the Middle English period, the Old English ambiguous adverb/​conjunctions þa and þonne, with distinct discourse-​pragmatic functions, and the interrogative hwonne develop into the Present-​Day English adverb then, which has retained some of the earlier discourse-​pragmatic functions, and the interrogative/​conjunction when. Earlier research (Kivimaa 1966; Mitchell 1965; Wårvik 1995; Yamakawa 1969) has found that when first replaces the conjunction þonne starting from South, South-​West and West Midlands, and it may even occur in the same texts as the conjunction tho, keeping their old discourse-​pragmatic distinctions in narrative structuring.2 Example (1) from the late thirteenth century South English Legendary illustrates this distinction. Þo ‘when’ (l. 139) refers to a unique event, while when (l. 141) refers to a repeated event. 1)

þo hi come hom to his In: he made hom faire chere & cortesliche serueþe hom: quinte man as he were. ac euer whan he from hom turneþe: þe teres ourne of his eie; (Horstmann 1887: 396–​397) ‘When they came home to his inn: he made them a nice feast /​and kindly served them: the courteous man that he was. /​but whenever (lit. always when) he turned from them: tears ran from his eyes;’

The adverbs merge into then, though they maintain their formal and functional distinction longer than the conjunctions, and tho in the sense ‘at that time’ still occurs dialectally in the South-​West (OED s.v. tho). In example (2) from the Chronicle of Robert of Gloucester (Wright 1887; henceforth RobGlo), þo ‘then’ (l. 8896) refers to a turn in a dialogue, i.e. a unique event following a previous event in temporal succession, whereas þanne ‘then’ (l. 8898) accompanies a question, referring to a future, and hypothetical, situation. (2)

Sire quaþ þis maide þo . þat is a vair name . As wo seiþ al is lif. & of grete fame .

2 The regular development of Old English þa gives us þo, tho, while the longer Old English item appeared as þonne, þanne, þenne, eventually merging on then (OED, s.v. then, s.v. tho; MED, s.v. thanne (adv.), s.v. tho (adv.)). When arises through multiple variant forms from Old English hwonne, hwanne, hwenne (OED, s.v. when; MED, s.v. whanne (adv. & conj.)). The variants found in the sample are þo, tho, than, thanne, thenne, þan, þanne, þen, wan, wanne, when, whenne, whan, and whanne. The choice of the forms tho and then for the adverbs and when for the conjunction to refer to Middle English data is motivated by convenience of reference rather than frequency.

The Demise of Ambiguous Adverb/Conjunctions

247

Ac wat ssolde is sone hote þanne . & oþere þat of him come3 (RobGlo 8896–​8898) ‘Sir, said this maiden then . that is a fair name . /​As they say all his life . and of great fame . /​But what should his son be called then . & others that of him come’

The aim of this paper is to study what different versions of the same text can reveal about the uses of these changing items in the same textual context. The paper reports a case study on manuscript variation in the uses of tho, then, and when in the text known as the Chronicle of Robert of Gloucester. The text is a long verse narrative in the West-​Midland dialect, the area where the temporal conjunction use of when emerged early and tho as an adverb and conjunction survived longest. The manuscripts in Wright’s edition (1887) are from early fourteenth to mid-​fifteenth century, the period when the latest instances of tho can be found in the Helsinki Corpus data (Wårvik, forthcoming).

2. Background While the main lines of development of the Present-​Day English temporal connective when are fairly straightforward, the detailed steps are still unclear. In Old English, its etymological ancestor hwonne was an interrogative introducing direct and indirect questions, while the most common temporal connectives corresponding to Present-​Day English when were þa and þonne, aptly called ambiguous adverb/​conjunctions by Mitchell (1985: § 2444, § 2536). These two items were not distinguished by their syntactic function, corresponding to both Present-​Day English when and then, but they had different discourse-​pragmatic functions: þa occurred in narrative texts marking the foregrounded main line and þonne occurred in all kinds of texts referring to hypothetical and habitual situations, as well as backgrounded material in narratives (Mitchell 1985: § 2562; Ogura 1984; Wårvik 2011a; cf. Bosworth –​Toller, s.v. þanne; on grounding see Wårvik 2011b). Mitchell notes that þa and þonne kept their complementary distribution all through the Old English period (1985: § 1116–​1117) and hwonne was not yet used to introduce temporal clauses, though it had already started to acquire that function, as evidenced by borderline cases between interrogatives and adverbial and relative clauses (1965, 1985: § 2779). The early Middle English developments of temporal conjunctions are charted by Kivimaa (1966) in her study of þe and þat as connectives in twelfth and thirteenth century texts. She observes that the Old English uses of the conjunctions continue in early Middle English, but when appears in the contexts of þonne in

3 Line numbers follow Wright’s edition (1887).

248

Brita Wårvik

the twelfth century and there are also occasional instances of þonne and hwanne in the contexts where one would expect þa. In the thirteenth century, when gets more frequent, replacing mainly the conjunction þonne. Changes in the uses of tenses reflect changes in the discourse functions of the items: with þa/​þo (and ase) the tenses of the past are used, with þonne those of the present and the future, but with their retreat, any tense may occur with them; in texts in which hwanne is infrequent, the tense is generally the present or the future, but if hwanne is common, the tense of the verb is the one required by the context. (Kivimaa 1966: 246)

Yamakawa, studying the development of conjunctions meaning ‘when’ until the end of the Middle English period, sees a slightly different beginning in the changes and notes “that first þanne advanced into the domain of þo where a preterite verb is appropriate, and then or rather at the same time hwanne began to supplant þanne in the new domain, while retaining the old one, originally occupied by þanne, where a verb denoting the present or future is appropriate” (Yamakawa 1969: 21). However, the majority of Yamakawa’s examples show the conjunctions in their original contexts and the rare instances illustrating the advance of þonne can hardly be seen as representing a general trend.4 Wårvik’s (1995) study of samples of saints’ lives from Old to late Middle English revealed similar trends for the conjunctions but suggested a stronger role for discourse factors. While in the Old English saints’ lives by Ælfric the temporal conjunctions þa and þonne are used, the longer item does not appear as a conjunction in the Middle English data. In the late thirteenth-​century South English Legendary, both tho and when appear as conjunctions, tho in the discourse contexts of Old English þa, and when in those occupied earlier by þonne. In Wårvik’s later samples, Mirk’s Festial and Caxton’s Golden Legend, from the late fifteenth century, when has taken over. Wårvik suggests that the discourse-​ structuring function as a signal of the main line of narrative may be a motivation for the survival of the conjunction tho alongside when in Middle English, while then only continued in use as an adverb. Both Kivimaa’s and Yamakawa’s studies reveal dialectal differences in the developments of these conjunctions. When as a temporal conjunction emerges

4 Of the two cases of þonne referring to singular events in the past, one is a variant in one of two manuscripts (Yamakawa 1969: 23, ­examples 13 and 13a, Lawman’s Brut) and the other occurs in a text with a majority of then-​forms (Yamakawa 1969: 23, ­example 11, Havelock the Dane). The single exceptional þa in the context of þonne involves þa gyt and may rather be an adverbial use ‘still then’ (Yamakawa 1969: 24, note 56; Lambeth Homilies).

The Demise of Ambiguous Adverb/Conjunctions

249

first in texts ascribed to the Southern and West-​Midland dialects (Kivimaa 1966; Yamakawa 1969), and uses of tho can still be found in texts from the South and South-​West in the fourteenth century (Yamakawa 1969). The use of tho alongside when in the South-​English Legendary, mentioned earlier, tallies with this distribution (Wårvik 1995). In the Helsinki Corpus, the latest cases of tho as adverb and conjunction are in the West and East Midlands texts in the period after 1350 (Wårvik, forthcoming).

3. Data The data for this study consist of a sample of 1,687 lines from the Middle English text called the Chronicle of Robert of Gloucester. So far fourteen different manuscripts have been identified of this text, the earliest dated to the early fourteenth and the latest to the sixteenth century (Hudson 1966; Pickering 2001; Shaw 2011; eLALME).5 Despite the traditional title given to the text in many sources, the identification of the author dates from the sixteenth century antiquaries, and the Robert referred to in the text (ll. 11748–​11749) may have been only the scribe of two of the manuscripts and even his connection with Gloucester is uncertain (Hudson 1969; Shaw 2011). Though the text has undoubtedly been popular, as shown by the number of manuscripts and later interest by antiquaries and anthologists (Edwards –​Gillespie 2003; Hudson 1969; Shaw 2012), its textual history has not yet been fully explored and the best edition available is still Wright’s edition from 1887. Wright’s edition includes eight of the fourteen manuscripts, dating from the early fourteenth to the mid-​fifteenth century (Table 1). Three of them belong to the so-​called first recension, or longer version, and five to the second recension, or shorter version (cf. Hudson 1966, 1969; Shaw 2012; Wright (ed.) 1887: xl–​ xlvi). The sample used in this study consists of the 1,687 lines shared by all these eight manuscripts: lines 4684–​4931; 4936–​5547; 5550–​5739; 5742–​5811; 8557–​ 8626; 8628–​8897; 8902–​8898; 8902–​8908; 8912–​8931; 8934–​9092; 9094–​9135.

5 The OED gives the manuscript dates as c1325–​c1425, but the dates of the quotations vary from 1297 to 1325 (>1300).

250

Brita Wårvik

Table 1:  Chronicle of Robert of Gloucester: Manuscripts in Wright’s edition Manuscript

Wright’s/​ Hudson’s sigla A /​C

London, British Library, Cotton Caligula A. xi. London, British C /​A Library Add. MS. 19677 London, British B /​H Library, Harley 201

Cambridge, Trinity College, R.4.26

α /​T

Oxford, Bodleian β /​D Library Digby 205 Cambridge, ε /​M Magdalene College, Pepys 2014 London University δ /​L Library 278 Cambridge ɣ /​U University Library Ee.4.31

Date LALME linguistic profile (LALME/​MED/​ Hudson) (county) c1325/​ c1325(c1300)/​ s. xiv in. c1400/​ c1400/​ s. xiv/​xv. c1400/​ c1425/​ s. xvi

7100 (Gloucestershire) 5300 (Wiltshire)

7500 (lines 1–​3177 Hand A Herefordshire) 7080 (lines 3178–​Hands B, C and D Gloucestershire) Last quarter of fourteenth 7070 century/​ (Gloucestershire) c1400(c1325)/​ c1400 Early fifteenth century/​ 8110 ?c1425/​ (Warwickshire) s. xv in. First quarter of fifteenth 4710 century/​ (Northamptonshire) c1425/​ s. xv in. -​/​ -​ c1450/​ s. xv in. -​/​ -​ c1450/​ s. xv mid.

Even if Wright notes that “[to] the student of the earlier forms of English the language of the Chronicle is of the greatest value” (1887: lx), the text has rarely been used as data for the history of language and many questions are still open or unexplored in its textual history. Shaw’s (2012) overview of earlier studies shows that the text has typically served as material for studying history, real, legendary or literary, while literary scholars seem to have adopted Wright’s view that “[as] literature, it is as worthless as twelve thousand lines of verse without

The Demise of Ambiguous Adverb/Conjunctions

251

any spark of poetry can be” (1887: lx). Apart from the linguistic profiles of six of the manuscripts in eLALME, the linguistic properties of the various versions have attracted little attention. A remarkable exception is Hudson’s edition of part of the chronicle (1964) and her study (1966) of the distribution of a number of traditional and innovative forms in the different manuscripts, including forms of third person pronouns and verbal inflections. Comparing the choices of forms, Hudson characterizes “the good manuscripts”, i.e. A, B and C of the first recension and α, δ and Q (Huntington Library, San Marino, California, HM 126) of the second recension, as “present[ing] orthographically, in varying degrees according to their dates, a remarkable state of conservatism” (1964: 311). In contrast to them, “both D [Wright’s β] and U [Wright’s γ] stand at the furthest remove from the original in the second recension” (1964: 80). These two manuscripts appear to stand apart from the rest also dialectally, displaying traces of eastern features (1964: 307–​ 308). Hudson’s findings of scribal practices lead her to conclude that the scribes were consistent in their choices: “except in those cases where misunderstanding can reasonably be postulated [...], a scribe will only copy the linguistic forms of his exemplar when they form part of his orthographic ‘vocabulary’ ” (1966: 371–​ 372). However, she (1966: 368–​369) observes that scribes treat rhyming positions differently in that they, particularly the scribe of β (1964: 309), may there preserve deviant forms which they elsewhere in the text change to forms more consistent with their idiolect. Another aspect of scribal practices is investigated by Kooper (2013), who examines ‘content markers’, i.e. devices signalling the organization of the text, in twelve manuscripts of RobGlo, including one prose version. He finds that the use of large capitals to indicate beginnings of sections in the narrative is fairly consistent and may even go back to the earliest manuscript (i.e. A), but there is great variation in where headers and marginal comments occur, which suggests that they are later additions. Comparing the manuscripts, Kooper singles out β as the manuscript with all the three types of markers. Though some of them are later additions, their use by the original scribe suggests a consciousness of text structure and a desire to indicate textual organization clearly. In his study of the development of when, Yamakawa (1969) comments on the use of tho and when in a sample from the Chronicle comprising 622 lines from the Caligula manuscript (A in Wright, C in Hudson) included in Skeat’s collection Specimens of Early English (1922). The contextual distribution of the items in this manuscript follows their original division of labour: there are 25 cases of þo in clauses with past or past perfect verbs and seven cases of when in clauses with past habitual or repeated events, of which one is a past subjunctive.

252

Brita Wårvik

The Linguistic Atlas of Late Mediaeval English includes descriptions of the language of six of the eight manuscripts in Wright’s edition (eLALME). The linguistic profiles include variants of then as item 30 and those of when as item 56, as presented in Table 2. According to the profiles, A and C use the wh-​ and þ-​ forms of the conjunction equally, while B and α prefer þ-​forms and β and ε the wh-​form. The dates of the manuscripts thus do not seem to go hand in hand with an increase in the newer wh-​form of the conjunction (cf. Table 1). Table 2:  Variants of THEN and WHEN in eLALMEa Manuscript A Date c1325

B c1400

C c1400

LP 30: THEN

7500/​7080 þo (þenne) ((þanne))/​ þo ((þanne, þo-​þat, þanne)) þo (wan, wen) ((þe, when, whan))/​ þo ((wan, wanne, wanne, þo-​þt))

5300 þo (þanne, þan)

7100 þo (þanne)

56: WHEN wanne, þo, ȝwanne (wan)

Α last quarter of fourteenth century 7070 þo (þanne)

whanne, þo, þo whenne (whanne, whan)

β early fifteenth century

Ε first quarter of fifteenth century

8110 þan, þanne (þenne)

4710 tho, þo, þan, than (þen, then)

when, whenne, whan, whanne ((tho))

whan (when, tho) ((whan))

a  The main forms appear without parentheses. Single parentheses indicate a frequency of occurrence one to two thirds of the main form, and double parentheses a frequency lower than one third of the main form (eLALME, “Introductions”, Vol. 3 “Linguistic Profiles”, section 13).

As the spread of when to replace the þ-​forms is said to have started from the South and West-​Midlands, it can be useful to see where the manuscript versions are assumed to have been written, even if this study does not concern dialects. Figure 1 shows the locations on the map of the six manuscripts that are included in LALME (eLALME).

The Demise of Ambiguous Adverb/Conjunctions

253

Figure 1:  Location of six manuscripts of Chronicle of Robert of Gloucester according to LALME linguistic profiles (LP)

Quite interestingly, assuming that wh-​ forms are more common where the change from þ-​forms started, there seems to be something of a mismatch between the locations and the preferences for þ-​ and wh-​forms according to the linguistic profiles of the manuscripts: B and α, where the older þ-​forms are more common, are located furthest west, while β and ε, which prefer the incoming wh-​ forms, are placed furthest east and among the northernmost ones in LALME.

4. Results In the sample of 1,687 lines of RobGlo, there is a total of 243 cases where at least one of the eight manuscripts uses tho, then, or when. Table 3a presents the alternatives used in the 144 instances of adverbial uses and Table 3b the forms in the 99 cases of conjunctive uses. The division into adverbial and conjunctive uses is based on the first recension manuscripts, which are followed in the second recension with only four exceptions where clause status is changed (see note b in Table 3a and 3b).

254

Brita Wårvik

Table 3a:  Alternative forms in cases of adverbial uses in the sample of RobGloa Item/​MS tho then 0 other Total

A 89 3 41 11 144

B 89 4 41 10 144

C 87 3 43 11 144

Α 95 4 33 12 144

Β 74 34b 22 14 144

ε 98 5 31 10 144

Δ 99 4 32 9 144

γ 89 5 39 11 144

all manuscripts 59 2 -​ -​ 61

a  For alternative forms see note 2 above. Cases with no adverbial or a gap in the text are subsumed under 0. Category “other” includes also, and, anon, mo, siþe/​siþþe/​suþþe, so, tuenty ȝer, þat, þat tyme, þe, þer/​þere, when. b  This excludes three cases where the β scribe rewrites a subclause introduced by tho as a main clause with then (cf. Table 6b). These cases are included in the category ‘other’ in Table 3b.

Table 3b:  Alternative forms in the cases of conjunctive uses in the sample of RobGloa Item/​MS tho when 0 other Total

A 72 15 4 8 99

B 73 17 4 5 99

C 72 16 5 6 99

α 75 16 4 4 99

β 53 32b 5 9 99

ε 74 17 3 5 99

δ 69 16 6 8 99

γ 69 18 4 8 99

all manuscripts 46 14 -​ -​ 60

a  For alternative forms see note 2 above. Cases with a gap in the text are classified as 0. Category “other” includes ar, as, sith/​siþe/​siþþe/​suþþe, so/​so that/​so þat, that/​þat, þen/​þan, þe wule, þus. b  This excludes one case where the β scribe rewrites a main clause with tho as a subclause introduced by when. This case is included in the category ‘other’ in Table 3a.

With one exception, the manuscripts largely agree in the proportions of the different alternatives in both adverbial and conjunction cases. Despite its potential syntactic ambiguity, the clearly preferred alternative is tho, which is used in 49–​70 percent of adverbial cases and 55–​75 percent of conjunction cases. Then only appears as an adverb, though much less frequently than tho, and the conjunction when is used in 15–​32 percent of the cases. The only manuscript that deviates from the general trend is β, which earlier studies have shown to be different even in other ways (cf. Hudson 1966; Kooper 2013). The figures in Table 3a and 3b mark this manuscript as more advanced in the disappearance of the ambiguous adverb/​conjunctions: the β scribe uses higher proportions of then as an adverb and when as a conjunction than the others. As was also noted earlier, the dates and locations of the manuscripts in LALME do not match the

The Demise of Ambiguous Adverb/Conjunctions

255

spread of the unambiguous forms from the South and West-​Midlands, and the distribution of the items in the sample of RobGlo is in line with that mismatch. There is, however, a slightly puzzling disagreement between LALME’s linguistic profile of β and the distributions in the sample: the linguistic profile 8110 appears to give a smaller role to tho, which, as a conjunction, is listed as occurring at a frequency less than one third of that of when and is not listed at all as an adverb. This raises the question of internal variation in this manuscript, even though it is unlikely that the samples of β in LALME would have been so different from those in this study. However, that investigation is outside the scope of the present study, and we now turn to considering the contexts of manuscript variation in the use of tho, then and when. Starting with instances of total agreement, we find that in 121, or about half of the 243 cases of tho, then and when, all the eight manuscripts agree. There is greater agreement in the use of conjunctions than adverbs: in 61 percent of the cases all manuscripts display identical uses of the conjunctions tho and when, whereas the adverbs tho and then are used similarly in only 42 percent of the cases. In all the instances of agreement the use of the items follows the Old English division of labour. On the one hand, tho as an adverb in the sense of ‘after that’ or ‘at that time’ and as a conjunction is used for single events, just like Old English þa, as in (3) and (4). (3)

(4)

þe erchebissop of kanterbury . anselin þat was þo . Fleu out of engelond . vor is vnriƷt and wo. (RobGlo 8618–​8619) ‘The archbishop of Canterbury . Anselm that was then./​Fled out of England . for his injustice and woe .’ þo king egbriȝt adde ybe king . seuene & þritti ȝer . He deyde as a noble mon . ac is londes he delde er . Adelwolf is eldore sone . he ȝef þo al west sex . (RobGlo 5202–​5204) ‘When king Edgbright had been king . seven and thirty years. /​He died as a noble man . but his lands he divided before. /​Adelwolf his elder son . he gave then all West Sussex .’

On the other hand, the two instances of then where all manuscripts agree, occur in the contexts of Old English þonne: once in an if-​then construction (5) and once in a description (l. 9003). (5)

ȝif it so is quaþ þe king . þat hii biddeþ vor vre fon. þanne viȝteþ hii aȝen vs . as moche as is hor miȝte. (RobGlo 4852–​4853) ‘If it is so, said the king, that they pray for our foes. /​Then they fight against us. as much as is their might .’

256

Brita Wårvik

Finally, the contexts of agreement tally with earlier findings of when first replacing the longer item: these clauses introduced by when depict repeated or hypothetical situations, i.e. those belonging to the domain of Old English þonne, as in (6). (6) Hire maidens brotte hire clene water . euere wanne heo hete6. (RobGlo 8965) ‘Her maids brought her clean water . whenever she commanded .’ Considering variation in the three first recension manuscripts, we find, as expected, more agreement, namely in 86 per cent of all cases. Table 4 shows the distribution of the alternatives in the cases where the first recension manuscripts agree with each other and disagree with one or more of the second recension manuscripts. For the uses of tho, the disagreements between the manuscript groups involve about a third of the cases. For when and then, which are much less frequent, there is only one instance of each where first recension manuscripts have chosen the same item in contrast to the second recension. In addition to these, in one instance all first recension manuscripts have there, while one second recension manuscript has tho and another one has then. Table 4:  Agreement among first recension manuscripts in contrast to second recension manuscripts in the uses of tho, then and when in the sample of RobGlo.a Item 0 then there tho when Total

adverb 38 1 1 23 -​ 63

conjunction 3 -​ -​ 21 1 25

Total 41 1 1 44 1 88

a  For alternative forms see note 2 above. Cases with a gap in the text are classified as 0.

In the second recension manuscripts there is no difference between conjunctions and adverbs in the proportions of the cases of agreement, both agreeing in 86 per cent of cases. However, if we ignore the cases where one or more second recension manuscripts add an item where there is no corresponding item in the first recension manuscripts, i.e. category 0, the proportion of agreement for adverbs decreases to 60 per cent, which is lower than for conjunctions, but

6 Wright has lete, which is the choice in A and γ, but that hardly makes any sense in this context.

257

The Demise of Ambiguous Adverb/Conjunctions

still higher than the proportion of agreement in all manuscripts. We will next turn to the kinds of items that are added, but it is worth noting that it is somewhat unexpected that adverbial expressions, mostly temporal, are added so often in the second recension manuscripts, which are supposed to be later and could thus be expected to use fewer temporal expressions in the narrative (on changes in uses of narrative discourse markers, see Brinton 1993, 1996; Fludernik 2000; Wårvik 1990a, 1995). Looking then at the kinds of items added in the second recension manuscripts in places where the first recension manuscripts agree on having no corresponding item (Tables 5a and 5b), we find that in most cases there is equally a gap, but we also find that the added items reveal the same trend as the overall preferences in the uses of tho, then and when: all manuscripts prefer tho as an adverb and as a conjunction, except β, which favours then and when. A further point worth noting about the cases where an item is added to the text is that there are no cases where the second recension manuscripts agree in their choices of the added items. Table 5a:  Items in second recension manuscripts where first recension manuscripts have a gap in an adverbial positiona Item/​MS 0 tho then so Total

α 25 10 2 1 38

β 14 6 17 1 38

ε 22 14 2 0 38

δ 22 14 2 0 38

Γ 33 4 1 0 38

a  For alternative forms see note 2 above. Cases with a gap in the text are classified as 0.

Table 5b:  Items in second recension manuscripts where first recension manuscripts have a gap instead of a conjunctiona Item/​MS 0 tho then when Total

Α 2 1 -​ -​ 3

β 2 -​ -​ 1 3

Ε 2 1 -​ -​ 3

δ 2 1 -​ -​ 3

Γ 3 -​ -​ -​ 3

a  For alternative forms see note 2 above. Cases with a gap in the text are classified as 0.

258

Brita Wårvik

In the same way, there is no agreement in the choices of items in the second recension manuscripts in the cases where all first recension manuscripts use tho (Tables 6a and 6b). However, we still find the same preferences: β tends to choose then and when, while the other second recension manuscripts mostly retain tho both as an adverb and as a conjunction. The one instance of when in Table 6a and three instances of then in Table 6b are cases where β rewrites the text changing the clause status (see note b in Table 3a and 3b). Table 6a:  Items in second recension manuscripts where first recension manuscripts use tho as an adverba Item/​MS 0 tho then there and when that time mo Total

α 2 20 0 1 0 0 0 0 23

Β 2 8 9 1 1 1 1 0 23

Ε 2 17 0 2 0 0 1 1 23

δ 4 18 0 1 0 0 0 0 23

Γ 2 19 1 1 0 0 0 0 23

a  For alternative forms of tho, then and when see note 2 above. 0 indicates omission or rephrasing.

Table 6b:  Items in second recension manuscripts where first recension manuscripts use tho as a conjunctiona Item/​MS 0 tho that when then and for Total

α 1 18 1 0 0 1 0 21

β 2 5 1 9 3 1 0 21

Ε 1 17 0 2 0 0 1 21

δ 2 16 2 0 0 1 0 21

Γ 0 19 1 1 0 0 0 21

a  For alternative forms of tho and when see note 2 above. 0 indicates omission or rephrasing.

The Demise of Ambiguous Adverb/Conjunctions

259

Turning finally to the least frequent items among the cases where the first recension manuscripts agree and differ from the second recension, we again detect the discourse patterns inherited from Old English. The one instance of then in Table 4 involves an omission in a hypothetical clause in direct speech, and in the one case of when, one manuscript uses siþ ’since’ to introduce a subclause with clearly causal sense (7). Both of these belong to the discourse-​functional domain of Old English þonne. (7)

þe bissopes him ansuerede . & þe abbot dinoc . Al wiþ grete reysons . & wit of hor boc . þat wanne [ε: siþ] hii adde hom sulue . erchebissop & king . Hii ne ssolde to englissemen . abuye riȝt no þing .(RobGlo 4817–​4820) ‘The bishops answered to him . and the abbot Dinoc. /​All with great reasons . and wit from their book. /​That when [ε: since] they had themselves . archbishop and king. /​They should not to the Englishmen . bow in absolutely no thing .’

As this text is verse, one factor that is likely to affect the choice of expressions is rhyme. There are 33 cases of the items in rhyming positions. In 20 of them, tho is used in all manuscripts, as in example (3) above; these line-​final cases account for one third of the 59 instances where tho appears in all of the manuscripts. In 6 line-​final cases tho appears in 1–​2 manuscripts, while the others have a gap; again, β stands out as it uses tho in 5 of these cases, while γ uses tho in 3, ε in 2, A in one case, and the four other manuscripts have no corresponding item. Other line-​ final variants of tho are two cases of mo (in 7 and 1 manuscript), two of also (in 7 and 6 manuscripts), so (in 7 manuscripts), twenty ȝer (in 7 manuscripts), and one case of a gap in one manuscript and then in one manuscript, which unsurprisingly is β. Thus, as Hudson (1966) observed, the rhyme position is fairly stable, and we can add that again β differs in its use of then line-​finally. Though the manuscripts show variation in the choices of items, there nevertheless appears a clear pattern in the contexts where the alternative forms are used. As in the cases of agreement between the manuscripts, tho appears, with few exceptions, as both adverb and conjunction, in clauses depicting single events, while then is used for referring to hypothetical or future events and when to repeated events or causes, in all manuscripts except β. Of a total of 34 instances where then is used by only one scribe, 29 are in β, in addition to which there are three cases where the β scribe uses then when rewriting a subclause as a main clause (see note b in Table 3a). Of the 19 occurrences of when as variant in only one manuscript, 15 are in β, to which can be added one case of a main clause rewritten as a subclause in this manuscript (see note b in Tabl 3b). In these cases, both then and when are used in clauses depicting single events, thus accounting

260

Brita Wårvik

for almost all the instances in those categories in Table 7a and 7b. In contrast, as was observed earlier, the cases where all manuscripts use when do not depict single events, and the same preference is evident in the few cases where β agrees with some other manuscripts about the use of when. There are two cases where seven manuscripts have when; in one of them ε has selected siþ in a causal subclause (­example 7) and in another A has opted for þat when depicting a repeated event. In a third case, β and ε have chosen when instead of þo in the other manuscripts to introduce a subclause depicting a single event. Table 7a:  Uses of the adverbs tho and then excluding cases of agreement in the sample of RobGlo Item tho then

Type single event repeated event single event hypothetical or future event

A 30

B 30

C 28

α 36

Β 15

ε 39 3

2

1

1 1

32

1

δ 39 1 2

γ 30 2 1

Table 7b:  Uses of the conjunctions tho and when excluding cases of agreement in the sample of RobGlo Item tho when

Type single event comment single event repeated event cause comment

A 25 1

B 26 1

C 25 1

α 27 2

1

2 1

1 1

1 1

Β 7 1 16 1 1

ε 26 2 2 1

δ 23 1 1 1

γ 22 1 1 1 1 1

The few deviations from the traditional discourse patterns of usage in the other manuscripts are all to some extent borderline cases. The one instance of the adverb tho referring to a repeated event is the line-​final occurrence, mentioned above, instead of mo in the other manuscripts (l. 5337). As to conjunction use, the exceptional commenting subclauses introduced by tho in fact comment on unique situations (­example 8 and l. 5560: þat in A, B, C, and β; tho in α, ε, and δ; when in γ), and the when-​clause depicting a single event in γ (­example 9) and one of those in ε (l. 8799) provide backgrounded comments on the main line of

The Demise of Ambiguous Adverb/Conjunctions

261

the narrative, while the other such when-​clause in ε is introduced by tho when (l. 9049). (8)

(9)

He let grede þoru al þe lond. þe vnri3te lawes vndo. Þat is fader & is broþer helde. & prisons deliueri al so. Þer was loye in to al þat lond. þo [δ: þat] men mi3te abbe lawe. (RobGlo 8726–​8728) ‘He let proclaim through all the land. the unjust laws to undo. /​That his father and his brother held. & prisoners delivered also. /​There was joy in all that land. when [δ: that] people could have law’ King henry þes tueye studes. verst bulde iwis. Suþþe [ε: þan] he weddede anoþer wif. þo [γ: when] ido was al þis. (RobGlo 9048–​9049) ‘King Henry these two places. first built indeed. /​After that [ε: then] he married another woman. when done was all this.’

To consider alternative choices, let us focus on the variants in other manuscripts in cases where β has then or when. Excluding the 14 cases where all manuscripts use when, we have 18 instances of when in β. In 16 of them it appears as a variant of tho, which is the choice in seven cases in all the other manuscripts, the other choices varying from omission to when, as and ere; β is the only manuscript using when in 15 of these 16 cases. All these 16 depict single events, and we thus again find when in the traditional discourse context of tho. In the two remaining cases when is used in all but one manuscript, once in a causal clause (example (7)) and once to depict a repeated event (l. 5253), which both belong to the discourse-​functional domain of Old English þanne. Of the 34 cases of then in β only two are shared by all other manuscripts, while in 29 cases only β has then. In eight cases the other manuscripts have tho and in five β adds then where the others have no corresponding item. In total there are 30 cases where β has then where the others have either tho, a gap, or, in three cases, then. This leaves one case of there and tho and one with anon and tho as variants in the other manuscripts where β has then. As already noted, all these instances of then are in clauses depicting single events. All this supports the conclusion that in the idiolect of the scribe of β, both the conjunction when and the adverb then had advanced quite far into the domain of tho. One clue to the stage in the development of when from an interrogative to temporal adverbial conjunction is the position of the subclause: interrogative subclauses introduced by when follow their main clause, while adverbial clauses may be preposed, postposed and even medial. Mitchell (1965, 1985: § 2779) comments that some Old English postposed hwonne-​clauses are ambiguous between an interrogative and adverbial interpretation, suggesting a beginning of the development towards adverbial clauses. In the present sample, 53 of the

262

Brita Wårvik

99 subclauses are preposed, while 35 are postposed and 11 medial. Of the total of 36 cases of when as variant conjunction, three are placed in the middle, 17 are postposed and 16 are preposed (The one case of a main clause rewritten as a subclause introduced by when is excluded). Ignoring the medial instances, we find a clear preference for when-​clauses to be postposed and tho-​clauses to be preposed: both in cases where all manuscripts agree on using tho and in those where tho and when appear as variants, two thirds of the subclauses are preposed, whereas the subclauses are postposed in three quarters of the cases where all manuscripts use when. Table 8:  Position of subclauses introduced by tho and when variants tho in all manuscripts tho and when when in all manuscripts

Preposed 28 12 3

Postposed 14 6 9

Focussing on subclauses introduced by tho and when in the different manuscripts, we can detect a preference for tho-​clauses to be preposed and when-​clauses to be postposed, except in β, where slightly more when-​clauses are preposed than postposed (Table 9).7 Table 9:  Position of subclauses introduced by tho and when in the sample of RobGlo Item /​ Position

Tho preposed

postposed

When preposed

postposed

MS A B C α β ε δ γ

44 42 43 45 32 44 41 42

21 22 21 21 17 22 19 20

4 4 4 4 16 4 4 4

9 11 10 10 14 11 10 12

7 For all manuscripts except β, the Chi square test shows that the differences are significant, p < .01.

The Demise of Ambiguous Adverb/Conjunctions

263

This further supports the conclusion that the use of when as an adverbial conjunction was more developed in the idiolect of the β scribe.

5. Factors Both internal and external factors have been proposed as involved in the changes in these connective items. The general trend towards less polyfunctionality can be considered as an internal factor contributing to the development of these items, among many others. However, in this development the syntactic polyfunctionality of the ambiguous adverb/​conjunctions þa and þonne was resolved at the expense of the loss of their distinct discourse-​ pragmatic functions. We could even say that Present-​Day English then, and perhaps when as well, are discourse-​pragmatically polyfunctional, at least to the extent that they need to cover for all the remaining discourse structuring functions of their predecessors. Moreover, apart from the general trend away from polyfunctionality, it can be difficult to show why syntactic ambiguity became a problem in Middle English, when it had not been one in Old English (cf. e.g. Kivimaa 1966: 257; Mitchell 1984, 1985: § 1879, 2536–​2560). While many of the connectives appearing in Middle English are borrowed, either as loans or as calques (Kortmann 1998; Lenker 2010; Rissanen 2012), the role played by foreign influences in the demise of the ambiguous adverb/​ conjunctions is less obvious. Kivimaa characterizes Scandinavian influence on conjunctions in Middle English as “considerable” (1966: 256), but she does not include the when group among those affected. However, she concludes that “it seems that Latin, probably together with French, was partly responsible for the advance of hwanne in temporal clauses” (1966: 257), but she appears to base this conclusion mainly on the absence of other motivations for why þonne was so much earlier replaced by the wh-​form than other þ-​connectives. To return to RobGlo, a simple test of foreign influence is to compare the origin of the items where the manuscripts vary in their lexical choices. RobGlo is the 93rd most quoted source in the OED, with a total of 3403 quotations, of which 315 are first quotations of current words (OED, Advanced search for “Robert of Gloucester” in First Quotation). Of these first quotation words, the OED identifies 144 as of native origin, 140 as of French, 12 of Latin and one of Scandinavian origin. Table 10 shows how the manuscripts vary according to the origin of variant lexical forms in the sample of the present study.

264

Brita Wårvik

Tab: 10:  Lexical variation in the sample of RobGlo type/​origin obsolete native Scandinavian Romance alternatives omission total items

Manuscript A B 61 62 13 13 0 0 5 4

C 60 13 1 5

α 60 14 0 4

79

79

1 79

79

β 29 17 19 12 1 1 79

ε 35 16 8 18 1 1 79

Δ 52 14 7 6

γ 38 14 21 6

79

79

Interestingly the same manuscripts where both then and when are used for referring to single events (cf. Table 7a and 7b) show the lowest proportions of obsolete lexical items (β, γ and ε), the highest proportions of Scandinavian loans (β and γ), and the highest proportions of words of Romance origin (β and ε). The manuscript with all these features is, unsurprisingly β. Thus, there appears to be a link between the development of the three connectives and the increase of foreign elements as well as the low proportion of obsolete items in the lexis, but it is of course not possible to claim that there is any causality in that link. As to other changes taking place during the Middle English period, let us take two at random and examine the manuscripts’ use of contracted negatives and verbal prefixes in past participles.8 Both the Old English type of negative contraction (e.g. ne+is = nis)9 and the Germanic prefixes marking past participles (Old English ge-​developing into y-​/​-i​ )10 were gradually disappearing, but the old and the new types are still both found in the versions of RobGlo. 8 One reviewer suggested constituent order as a potential indicator of the progress of the development towards Present-​Day English uses of then and when. However, constituent order appears to vary only rarely in this sample, thus not providing enough cases for patterns to be detected. 9 In his study of negative contraction in Middle English, Hogg (2004) notes that the Old English type survived until mid-​fifteenth century but was then fast disappearing. He identifies Gloucestershire as a core area of this type of contractions, which motivates my choice of this feature as a comparative variable. 10 Hiltunen (1983:65) notes that the ge-​prefix tended to be characterized by “semantic and functional vagueness” already in Old English, but that tendency became apparent in Early Middle English. Pilch (1955) observes that the i-​prefix of the past participle disappears beginning from the North by the end of the fourteenth century. The variant

265

The Demise of Ambiguous Adverb/Conjunctions Table 11:  Variation in negative forms and uses of prefixes in the sample of RobGlo

contracted non-​contracted Negatives total prefixed without prefix Variant verbs total

Manuscript A B 75 71 1 5 76 76 190 190 8 8 198 198

C 73 3 76 191 7 198

α 74 2 76 187 11 198

β 13 63 76 104 94 198

ε 68 8 76 152 46 198

Δ 48 28 76 168 30 198

γ 67 9 76 91 107 198

In the choices of the negative forms, the clearly deviant manuscript is, again, β, while also δ has a somewhat higher proportion of non-​contracted forms. The uses of prefixes show γ to be further advanced in the disappearance of the prefixed forms, with β close behind. These variations tally with earlier judgements of α and to some extent even δ as conservative manuscripts, and β and γ as more progressive manuscripts among the second recension (Hudson 1964; cf. above).

6. Conclusions To sum up, the variations between the manuscripts of RobGlo reveal interesting differences in the scribes’ use of tho, then and when. Whereas the scribes in most cases follow the discourse-​pragmatic patterns inherited from Old English, the scribe of β deviates from the traditional usage in being more advanced in the adoption of the disambiguated items. Then no longer appears as a conjunction in any of these manuscripts; instead, the scribes use when and tho. While these conjunctions mainly keep to their traditional discourse contexts, when in the contexts of Old English þonne and tho in the contexts of Old English þa, in manuscript β, when also occurs in the contexts of tho-​conjunction. The adverbs tho and then likewise follow the traditional patterns, strictly in all other manuscripts except β, where then is frequently selected in contexts depicting single events, thus expanding into the discourse domain of Old English þa. In addition to being more progressive in the use of then and when, the β manuscript also stands apart in other respects. Hudson (1964) characterizes its

forms in Table 11 include cases where one or more scribe uses the prefix i-​/​y-​ and one or more a participle without the prefix. In 13 cases, some scribes use other prefixes (11 a-​, 1 for-​and 1 bi-​).

266

Brita Wårvik

orthography, and that of γ, as further away from the original and notes eastern dialectal features in these manuscripts. Kooper (2013) observes that β contains more content markers indicating textual organization. In the present study, β turned out to contain more preposed when-​clauses, fewer obsolete lexical items, more Scandinavian (after γ) and Romance (after ε) loans, fewer contracted negatives and fewer prefixed past participles (after γ) than other manuscripts. Though these factors constitute a somewhat random selection, they nevertheless all point to β being more progressive than the other manuscripts. Considering the dates of the manuscripts, this finding is not totally surprising, but it does not fully tally with the dates, because β is not judged to be the latest of the manuscripts (Table 1). Dialectal differences might provide an explanation, but, as when spread starting from the South-​West, its higher frequency in a manuscript with eastern features appears rather contradictory. In LALME’s linguistic profile maps, β is located north-​east among the manuscripts of RobGlo, which may be speculated to point to progressive tendencies related to northern influences. Though β appears progressive according to all these factors, there is one feature that goes against this characterization. When the development of narrative style towards more modern form involves a decrease in the use of explicit signals of the sequential story-​line, we would expect fewer instances of tho and then. What we find, however, is that the β scribe inserts then in places where there is a gap in other manuscripts, thus adding such signals. This may be related to the tendency observed by Kooper for the β scribe to use more content markers; both types of signals organize and structure the text. Thus, it could be argued that in this scribe’s production the need to mark discourse structure overrides the stylistic trend to use fewer explicit markers of story-​line. And the β scribe is not alone in this, nor is RobGlo the only Middle English text where we find items marking the main line of the story in a similar fashion to Old English þa (e.g. Malory’s Morte D’Arthur (Fludernik 2000) and saints’ lives (Wårvik 1995)). The changes in these ambiguous adverb/​ conjunctions coincide with a number of other developments transforming English from Old to Modern. Leaving aside the influx of loanwords and the drastic simplification of inflectional morphology, students of this period in the history of English still need to deal with changes that suggest typological shifts in syntax and discourse pragmatics. Such changes include the increasing frequencies of temporal subordinate clauses, decreasing frequencies of temporal adverbials showing the progress of the narrative story-​line, most notably þa, the disappearance of weorðan and its replacement by be (beon/​wesan), the emergence of the be+ V-​ing (beon/​ wesan+ V-​ende) and its development to the progressive aspect, and the word order changes from V2 (or a version of it) to SVO (Brinton 1993, 1996; Fischer

The Demise of Ambiguous Adverb/Conjunctions

267

1992; van Kemenade –​Los 2006; Los 2009; Petré 2010; Warner 2007; Wårvik 1995; among others). Though these changes may appear unrelated, they can be seen as parts of larger developments. Los (2012) relates some of them to a typological shift of English from a bounded to an unbounded construal in narrative structuring (e.g. Carroll et al. 2004). Along similar lines, Wårvik (1990b) characterized the changes in signalling of narrative structure during the Middle English period as a typological, rather than a stylistic shift from a foreground-​ marking language to a background-​ marking or fuzzy-​ grounding language. Whatever the interdependences between these almost parallel developments in the Middle English period are, it is difficult to see how their coincidence could be purely accidental. The findings of this present study of manuscript variation of one text tally with the hypotheses that the parallel changes are parts of larger developments taking place during the Middle English period. Obviously, more studies, particularly of texts with multiple manuscripts, are needed to determine how exactly the different changes are related.

References Adamson, Sylvia –​Vivien A. Law –​Nigel Vincent –​Susan Wright (eds.). 1990. Papers from the 5th International Conference of English Historical Linguistics. (Current Issues in Linguistic Theory 65). Amsterdam & Philadelphia: John Benjamins. Aertsen, Henk –​Robert J. Jeffers (eds.). 1993. Historical Linguistics 1989. Papers from the 9th International Conference on Historical Linguistics. (Current Issues in Linguistic Theory 106). Amsterdam & Philadelphia: John Benjamins. Andersen, Henning –​Konrad Koerner (eds.). 1990. Historical Linguistics 1987. Papers from the 8th International Conference on Historical Linguistics. (Current Issues in Linguistic Theory 66). Amsterdam & Philadelphia: John Benjamins. Blake, Norman (ed.). The Cambridge History of the English Language. Vol. 2: 1066–​1476. Cambridge: Cambridge University Press. Bosworth, Joseph –​T. Northcote Toller. 2010. An Anglo-​Saxon Dictionary Online. Compiled by Sean Christ and Ondřej Tichý. http://​www.bosworthtoller.com [access 31 December 2018] [print version 1954, 1st printing 1898, Supplement 1921. Oxford: Oxford University Press.] Brinton, Laurel J. 1993. “Episode boundary markers in Old English discourse”. In: Aertsen, Henk –​Robert J. Jeffers (eds.): 73–​89. Brinton, Laurel J. 1996. Pragmatic Markers in English: Grammaticalization and Discourse Functions. (Topics in English Linguistics 19). Berlin & New York: Mouton de Gruyter.

268

Brita Wårvik

Carroll, Mary –​Christiane von Stutterheim –​Ralph Nuese. 2004. “The language and thought debate: A psycholinguistic approach”. In: Pechmann, Thomas –​ Christopher Habel (eds.): 183–​218. Curzan, Anne –​Kimberley Emmons (eds.). 2004. Studies in the History of the English Language II: Unfolding Conversations. (Topics in English Linguistics 45). Berlin & New York: Mouton de Gruyter. Edwards, Anthony S. C. –​Alexandra Gillespie. 2003. “A manuscript of Robert of Gloucester’s Chronicle copied by John Stow”. Notes and Queries 50 (248): 384–​385. eLALME = Benskin, Michael –​Margaret Laing –​Vasilis Karaiskos –​Keith Williamson. 2013–​. An Electronic Version of A Linguistic Atlas of Late Mediaeval English. Edinburgh: University of Edinburgh. http://​www.lel.ed.ac. uk/​ihd/​elalme/​elalme.html [accessed 31 December 2018]. Fischer, Olga. 1992. “Syntax”. In: Blake, Norman (ed.): 207–​408. Fludernik, Monica. 2000. “Narrative discourse markers in Malory’s Morte D’Arthur”. Journal of Historical Pragmatics 1: 231–​262. Hegedűs, Irén –​Alexandra Fodor (eds.). 2012. English Historical Linguistics 2010. Selected Papers from the Sixteenth International Conference on English Historical Linguistics (ICEHL 16), Pécs, 23–​27 August 2010. (Current Issues in Linguistic Theory 325). Amsterdam & Philadelphia: John Benjamins. Hiltunen, Risto 1983. The Decline of Prefixes and the Beginnings of the English Phrasal Verb. The Evidence from Some Old and Early Middle Texts. (Annales Universitatis Turkuensis, Ser. B, Tom. 160). Turku: University of Turku. Hogg, Richard M. 2004. “The spread of negative contraction in Early English”. In: Curzan, Anne –​Kimberley Emmons (eds.): 459–​482. Horstmann, Carl (ed.) 1887. The Early South English Legendary. (EETS OS 87). London: Trübner. Hudson, Anne. 1964. An edition of part of the chronicle attributed to Robert of Gloucester with a study of the original language of the poem. Unpublished D.Phil. dissertation, University of Oxford. Hudson, Anne. 1966. “Tradition and innovation in some Middle English manuscripts”. Review of English Studies n.s. 17.68: 359–​72. Hudson, Anne. 1969. “Robert of Gloucester and the Antiquaries, 1550–​1800.” Notes and Queries 214 n.s. 16.9: 322–​333. Jucker, Andreas (ed.). Historical Pragmatics. Pragmatic developments in the history of English. (Pragmatics and Beyond New Series 35). Amsterdam & Philadelphia: John Benjamins.

The Demise of Ambiguous Adverb/Conjunctions

269

Kemenade, Ans van –​Bettelou Los. 2006. “Discourse adverbs and clausal syntax in Old and Middle English”. In: Kemenade, Ans van –​Bettelou Los (eds.): 224–​248. Kemenade, Ans van –​Bettelou Los (eds.). 2006. The Handbook of the History of English. Malden, Mass.: Blackwell. Kivimaa, Kirsti. 1966. Þe and Þat as Clause Connectives in Early Middle English with Especial Consideration of the Emergence of the Pleonastic þat. (Commentationes Humanarum Litterarum 39/​1). Helsinki: Societas Scientiarum Fennica. Kooper, Erik. 2013. “Content markers in the manuscripts of Robert of Gloucester’s Chronicle”. In: Kooper, Erik –​Sjoerd Levelt (eds.): 43–​74. Kooper, Erik –​Sjoerd Levelt (eds.). 2013. The Medieval Chronicle VIII. Amsterdam: Rodopi. Kortmann, Bernd. 1998. “The evolution of adverbial subordinators in Europe”. In: Schmid, Monika S. –​Jennifer R. Austin –​Dieter Stein (eds.): 213–​228. Lenker, Ursula. 2010. Argument and Rhetoric. Adverbial Connectors in the History of English. (Topics in English Linguistics 64). Berlin & New York: Walter de Gruyter. Los, Bettelou. 2009. “The consequences of the loss of verb-​ second in English: Information structure and syntax in interaction”. English Language and Linguistics 13: 97–​125. Los, Bettelou. 2012. “The loss of verb-​second and the switch from bounded to unbounded systems”. In: Meurman-​Solin, Anneli –​Maria-​José López-​ Couso –​Bettelou Los (eds.): 21–​46. MED = Kurath, Hans –​Sherman M. Kuhn et al. 1952–​2001. Middle English Dictionary. Ann Arbor, Michigan: University of Michigan Press. Online edition in McSparran, Frances et al. (eds.). 2000–​2018. Middle English Compendium. Ann Arbor, MI: University of Michigan Library. https://​quod.lib.umich.edu/​ m/​middle-​english-​dictionary/​dictionary [accessed 31 December 2018]. Meurman-​Solin, Anneli –​Ursula Lenker (eds.). 2011. Connectives in Synchrony and Diachrony in European Languages. (Studies in Variation, Contacts and Change in English 8). Helsinki: VARIENG. Meurman-​ Solin, Anneli –​Maria-​ José López-​ Couso –​Bettelou Los (eds.). 2012. Information Structure and Syntactic Change in the History of English. New York: Oxford University Press. Mitchell, Bruce. 1965. ‟The status of hwonne in Old English”. Neophilologus 49: 44–​57. Mitchell, Bruce. 1985. Old English Syntax. Vols. 1–​2. Oxford: Clarendon Press.

270

Brita Wårvik

OED = Oxford English Dictionary Online. Oxford: Oxford University Press. http://​www.oed.com [accessed 31 December 2018]. Ogura, Michiko. 1984. “OE temporal conjunctions denoting ‘when’ or ‘while’: With special regard to the Gospels and the Psalter”. Neuphilologische Mitteilungen 85: 273–​290. Pechmann, Thomas –​Christopher Habel (eds.). 2004. Multidisciplinary Approaches to Language Production. (Trends in Linguistics: Studies and Monographs 157). Berlin: Mouton de Gruyter. Petré, Peter. 2010. “The functions of weorðan and its loss in the past tense in Old and Middle English”. English Language and Linguistics 14(3): 457–​484. Pickering, O.S. 2001. “South English Legendary style in Robert of Gloucester’s Chronicle”. Medium Ævum 70: 1–​18. Pilch, Herbert. 1955. “Me. I-​beim Participium Präteriti”. Anglia 73: 279–​291. Rissanen, Matti. 2012. “Grammaticalization, contact and corpora”. In: Hegedűs, Irén –​Alexandra Fodor (eds.): 131–​151. Schmid, Monika S. –​Jennifer R. Austin –​Dieter Stein (eds.). 1998. Historical Linguistics 1997. Selected Papers from the 13th International Conference on Historical Linguistics, Düsseldorf, 10–​ 17 August 1997. (Current Issues in Linguistic Theory 164). Amsterdam & Philadelphia: John Benjamins. Shaw, Philip A. 2011. “Robert of Gloucester and the medieval chronicle”. Literature Compass 8/​10: 700–​709. Shaw, Philip. 2012. “The composition of the metrical chronicle attributed to Robert of Gloucester”. English Manuscript Studies: 1100–​700 17: 140–​54. Warner, Anthony. 2007. “Parameters of variation between verb-​subject and subject-​verb order in late Middle English”. English Language and Linguistics 11.1: 81–​111. Wårvik, Brita. 1990a. “On grounding in English narratives: A diachronic perspective”. In: Adamson, Sylvia –​Vivien A. Law –​Nigel Vincent –​Susan Wright (eds.): 559–​75. Wårvik, Brita. 1990b. “On the history of grounding markers in English narrative: Style or typology?”. In: Andersen, Henning –​Konrad Koerner (eds.): 531–​42. Wårvik, Brita. 1995. “The ambiguous adverbial/​conjunctions þa and þonne in Middle English: A discourse-​pragmatic study of then and when in Early English Saints’ Lives”. In: Jucker, Andreas (ed.): 345–​357. Wårvik, Brita. 2011a. “Connective or ‘disconnective’ discourse marker? Old English þa, multifunctionality and narrative structuring”. In: Meurman-​Solin,

The Demise of Ambiguous Adverb/Conjunctions

271

Anneli –​Ursula Lenker (eds.). http://​www.helsinki.fi/​varieng/​journal/​ volumes/​08/​warvik/​ [accessed 31 December 2018] Wårvik, Brita 2011b. “Grounding”. Handbook of Pragmatics, Volume 15 (2011), pp. 1–​16. https://​doi.org/​10.1075/​hop.15.gro1 [accessed 24 June 2020] Wårvik, Brita. Forthcoming. “A corpus study of the frequency, distribution, and functions of ‘then’-​words in the history of English”. Wright, William Aldis (ed.). 1887. The Metrical Chronicle of Robert of Gloucester. Rolls Series 86. 2 vols. London: Her Majesty’s Stationery Office. Yamakawa, Kikuo. 1969. “The development of when as subordinate conjunction or relative adverb”. Hitotsubashi Journal of Arts and Sciences 10: 8–​42.

Richard Zimmermann

The Loss of Negative Concord with Negative PPs1 Abstract: Middle English loses the negative particle ne in its function as a sentential negator as well as in negative concord structures. This paper measures ne-​drop in both contexts with quantitative data from three syntactically parsed Middle English corpora. Surprisingly, the results show that the rate of ne-​loss with negative prepositional phrases (PPs) seems to be significantly slower than elsewhere. Different explanations for this observation are explored. It is argued that the predominant reason for this differential behaviour lies in one particular type of negative PP –​those usually taking bare quantifier complements and denoting ‘nothingness.’ These phrases do not take clausal scope and hence never occur with ne. Future studies should control for the presence of this type of negative PP to avoid over-​inflation of the category of ne-​free negative clauses. Keywords: Middle English negation, negative concord, Jespersen Cycle, negative PPs, parsed corpora, regression modelling

1. Introduction: The Loss of Ne Middle English inherits the negative particle ne from Old English, in fact from ancestral language stages going back all the way to Proto-​Indo-​European (e.g., Malory –​Adams 2006: 422). During the Middle English period, however, this item is in decline and eventually disappears. It is possible to identify two distinct environments for this loss. First, the particle ne disappears in its function as sentential negation (SN). Specifically, Middle English executes the Jespersen Cycle, a gradual change in SN from a single negative particle, ne, (Stage I, (1a)), through co-​occurrence of ne and the negative adverb not, (Stage II, (1b)), towards exponence of not alone

1 Parts of this paper were presented at the International Conference on Middle English (ICOME 11), Florence, February 2019, Swiss Works in English Language and Linguistics (SWELL 2019), Lausanne, March 2019, and the York-​Holland Symposium on the History of English Syntax (SHES 17), Edinburgh, May 2019. I would like to thank everybody who commented on my work at these conferences. I would also like to thank three anonymous reviewers for their helpful feedback as well as the editors of the present volume.

274

Richard Zimmermann

(Stage III, (1c)) (e.g., Jespersen 1917, Mitchell 1985: § 1603, Frisch 1997, Wallage 2008, Wallage 2017: ­chapter 5). (1)

Disappearance of ne as SN a. he ne sculde beon ded he not should be dead ‘He should not be dead’ (Layamon’s Brut, c. 1200) b. He ne shal nouʒt sechen. he not shall not seek ‘He shall not seek’ (Earliest Prose Psalter, c. 1350) c. They Ø2 shall not mysse they shall not miss ‘They shall not miss’ (A right merrie Comedie, c. 1570)

Second, the particle ne also disappears from Negative Concord (NC) structures. In this paper, NC is defined as the co-​occurrence of the negative particle ne with another negative element other than not to express a single logical negation. For example, ne can co-​occur with a negative adverb, (2), a negatively quantified object (3), or a negatively coordinated subject, (4) (e.g., Jack 1978, Iyeiri 1999). It is then dropped, leaving a single negative element behind. The resulting structure lets negation from the remaining negative element scope over the entire clause in lieu of SN. This is still a grammatical option in Modern English (modulo other linguistic changes). (2)

Disappearance of ne in NC with negative adverbials a. And þo .ij. sterres ne meeven neuere and those two stars not move never ‘And those two stars never move’ (Mandeville’s Travels, c. 1371) b. for þey Ø synneden neuere for they sinned never ‘because they never sinned’ (Wycliffe Sermons, c. 1383)

2 The null symbol Ø highlights throughout this paper the absence of the negative particle ne. A theoretical analysis or commitment to an empty element in not thereby implied.

The Loss of Negative Concord with Negative PPs (3)

(4)

275

Disappearance of ne in NC with negative objects a. they no haveth [no joye] they not have no joy ‘They have no joy’ (Alisaunder, c. 1300) b. On hyme ʒe Ø had [no mercy] on him you had no mercy ‘You had no mercy on him’ (Siege of Jerusalem, c. 1500) Disappearance of ne in NC with negative subjects a. [cayser no kyng] No may withstonde Godis helpyng! Ceasar nor king not may withstand God’s helping ‘neither Caesar nor king may withstand God’s helping’ (Alisaunder, c. 1300) b. [no fur. ne hete] ney me Ø com no fire nor heat nigh me came ‘neither fire nor heat came near me’ (South English Legendary, c. 1265)

NC can also be formed with more than one negative element (not counting ne). I will refer to such structures as multiple negative concord (multiple NC). Once again, Middle English witnesses the gradual disappearance of ne from such constructions, as illustrated in (5) for multiple NC with both a negative adverb as well as a negatively coordinated subject. Cases of multiple NC without ne are therefore best thought of as remnants after ne-​drop has taken place.3

3 Scholars do not always conceptualise multiple NC without ne as a leftover from an original construction that did involve the negative particle ne. Instead, they treat the structure as a primitive of investigation in its own right, especially as a variant in the spread of negative polarity items (NPIs) roughly between the fifteenth and eighteenth century (e.g., Nevalainen 2006, Ingham 2006, Kallel 2007), as illustrated in (i). (i) Development after the loss of ne: From multiple NC without ne to NPIs a. [No man] make [none congregaciouns] no man make no congregations ‘Nobody should hold any congregations’ (Brembre I, 1384, Ingham 2006 ex. 6) b. [No mannes cursynge] hath [ony strengthe] no man’s cursing has any strength ‘Nobody’s cursing has any strength’ (Wyclif tracts, c. 1380, Ingham 2006 ex. 11) These scholars assign the label NC exclusively to structures such as (ia) or (4a), whereas the present paper views those as special cases of NC (i.e., multiple NC without ne) and

276 (5)

Richard Zimmermann Disappearance of ne in multiple NC a. [neither God ne man] ne sholde nevere knowe it. Neither God nor man not should never know it ‘Neither God nor man should ever know it’ (The Parson’s Tale, c. 1390) b. [nouther the kyng of Ermonye Ne the contree] Ø weren neuer in pees neither the king of Ermonye Nor the country were never in peace ‘Neither the king of Ermonye nor his country were ever at peace’ (Mandeville’s Travels, c. 1371)

There is a close relation between the drop of ne in SN and NC. Nevertheless, it makes sense to distinguish the two patterns. (a) Step I of Jespersen’s Cycle only makes sense for SN, not for NC. (b) Theoretical analyses of the two phenomena usually assign to them two distinct semantic representations, SN being a relatively simple operation inverting the truth conditions of a proposition, NC involving more complex semantic types (e.g., Zeijstra 2004: ­chapter 8). (c) SN does not develop negative polarity items (NPIs) as an additional variant of negation whereas NC does. (d) Finally, the patterns described by SN and NC hardly ever overlap in the sense that if ne co-​occurs with not it does not usually also co-​occur with another negative element (e.g., Laing 2002). This paper focuses on one particular form of Middle English NC that has so far received relatively little attention in the literature, namely NC with negative prepositional phrases (PPs). Unsurprisingly, ne is lost in this context, too. Illustrations are presented for negative adjunct PPs (i.e., elements intuitively classified as optional modifiers) in (6) and negative complement PPs (i.e., elements intuitively categorised as non-​optional arguments) in (7). (6)

Disappearance of ne in NC with negative PPs a. he ne mihte cysten þærof [PP for nan þing] twelfe penegas. he not might cost thereof for no thing twelve pennies ‘he could not even get twelve pennies for it in any way’ (Peterborough Chronicle, c. 1125)

uses the term NC more liberally for all occurrences of multiple negative elements other than ne … not, as in (2), (3), (4), (5) etc.

The Loss of Negative Concord with Negative PPs

(7)

277

b. & sche Ø wolde [PP in no wise] gon thens and she would in no wise go thence ‘and she would absolutely not depart from there’ (Book of Margery Kempe, c. 1438) a. þey ne recche [PP of no ryʒt fey] they not care of no right faith ‘They do not care about the right faith’ (Handling Sin, c. 1303) b. þei Ø recche [PP of no schame] they care of no shame ‘They do not care about shame’ (Temporale Sermon cycle, c. 1420)

It is possible to measure and quantify the speed with which ne is lost in the SN and NC patterns illustrated in (1)–​(7).When such an investigation is conducted, the results show a surprising divergence: The negative particle ne is seemingly lost substantially more slowly with negative PPs than elsewhere. Specifically, there are on average more negative PPs without the negative particle ne than would be expected given a particular time point. Section 2 presents the corpus-​ based study that leads to this observation. The remainder of the paper will then investigate reasons for this prima facie unexpected behaviour of NC with negative PPs. Specifically, Section 3 pursues three avenues towards potential explanations: measurement errors, the argument –​adjunct distinction, and semantic differences. Section 4 concludes.

2. A Measurement of the Loss of Ne This section reports the observation that the loss of ne in NC structures with negative PPs seems to proceed surprisingly slowly. First, the data collection process will be described. Subsequently, the resulting dataset will be described, and a mixed-​effects logistic regression model will be fit to the data as a convenient way to summarise the development. I will also offer some important implications of the findings for related research.

2.1. Data Collection The data for this study come from three syntactically parsed and annotated corpora of Middle English, the Penn-​Parsed Corpus of Middle English, 2nd edition (PPCME2, Kroch –​Taylor 2000), the Parsed Corpus of Middle English Poetry (PCMEP, Zimmermann 2015) and the Parsed Linguistic Atlas of Early Middle English (PLAEME, Truswell et al. 2019). In total, the data basis consists of 168 electronic text files comprising c. 1.5 million words of running texts in c. 120,000 sentence tokens. Every text file has been assigned a rough date of composition

278

Richard Zimmermann

based on arguments offered in the professional literature. Observations were made within single finite clauses. All examples will henceforth be cited by their respective corpus conventions. The loss of ne in SN, (1), has been quantified several times before and hence scholars have established good measurement procedures that can be replicated here (Frisch 1997: 33, Wallage 2008: 645, Table 1, Newberry et al. 2017: 225, Truswell et al. 2019: 30). The present quantification groups together all ne and ne… not structures (Stage I, Stage II) and compares them to clauses with not on its own (Stage III). It excludes all sentences with additional negative items, such as negative objects, etc. The operationalization of the negative particle ne and the secondary negator not is carried out in the following way: Items are classified as ne if they are labelled as NEG and do not include the grapheme t (e.g., ne, ni, no). Conversely, items are categorised as not if they are labelled NEG and do include t (e.g., nawhiht, noht, not) or, rarely, co-​occur with another NEG without t (e.g., ne ... na). Moreover, the ne counts include all instances of ne fused to a finite verb, such as nis ‘is not’ nat ‘does not know’ or nolde ‘would not’ (contractions of this kind are excluded e.g., in Wallage 2008), (8). (8)

Example of fused (contracted, clitic) ne –​counted towards ne alle ðe ðine Behode healden nelleð all that your Order hold not-​want ‘all who do not want to obey your order’ (CMVICES1,19.224)

Likewise, instances of expletive ne, required in some embedded contexts but not actually interpreted as semantically negative, are indiscriminately added to the ne dataset (e.g. Jack 1977, Wallage 2017: ­chapter 3), (9). (9)

Example of expletive (redundant, paratactic) ne –​counted towards ne Sche wold for noman wond [þat sche no wold to him fond] she would for no-​man hesitate that she not would to him proceed ‘She would hesitate for nobody to proceed to him’ rather than: # ‘She would hesitate for nobody not to proceed to him’ (AmisAmiloun,28.[Stanza_​46].552.223)

The secondary negator not is considered in its typical post-​verbal position, (1c), as well as in other, rarer syntactic configurations (position as a diagnostic of

The Loss of Negative Concord with Negative PPs

279

grammaticalization in Frisch 1997, analysis of syntactic positions of not e.g., in Haeberli –​Ingham 2007, van Kemenade 2011), (10). (10)

Example of atypical position of not –​counted towards not Ðe oðer him Ø awrecþ ðeruore naʒt the other not him avenges therefore ‘the other does not therefore avenge it’ (CMAYENBI,115.2210)

Some forms of negation are ambiguous. In particular, forms of not can often be interpreted as SN or as a quantifier ‘naught, nothing’ (11a). Another kind of ambiguity arises with initial ne, which could be parsed not only as SN but also as a negative conjunction ‘nor’ (11b). Cases such as these are not disambiguated manually but classified in accordance with the corpus annotation. (11)

Ambiguous forms of SN –​counted according to corpus annotation a. Spare Ø wold he nouʒht. spare would he not/​nothing i) He would not show mercy (spare). (= intransitive + negation) ii) He would forgive (spare) nothing. (=transitive + negative object) (SirCleges,46.210.115) b. Ne sal þeih no man samie þiar. ne þarf he him adrade. not shall though no man shame there not/​nor need he himself fear i) No man shall be ashamed there though. He shall not be afraid (= negation) ii) No man shall be ashamed there though. Nor shall he be afraid (= conjunction) (PoemaMorale,143–​144.225.165–​167)

The searches returned 9,354 examples of SN (4,800 with ne, 4,554 without ne). Next, the loss of ne is measured in NC structures with negative adverbs and adverbial phrases, (2). The most frequent lexical item in this group is the adverb never. However, several additional expressions are counted as well, including nevermore, nowhere, no more, no longer, nevereft, etc. An example is given in (12). As before, hits were only considered if they did not involve other negative elements, such as negative objects, negative subjects, etc.

280 (12)

Richard Zimmermann Additional example of loss of ne in NC with negative adverbial wikkidnesse a. thei ne were neveremo withouten the tormentz of hir they not were nevermore without

the torments of their

wickedness

‘they were nevermore without the torments of their wickedness’ (CMBOETH,448.C1.398)

b. The the

Jewes being in ther lyberte or captiuyte. Ø had Jews being in their liberty or captivity had

neuer moo kynges never more kings

‘The Jews, whether in liberty or captivity, never had kings again’ (CMFITZJA,A5R.68)

The search queries retrieved 1,881 relevant hits (868 with ne, 1,013 without ne). The next measurement concerns the loss of ne in NC with negative objects, (3), and negative subjects, (4). Elements are classified as negative when they are negatively quantified, (3), or negatively coordinated, (4). Negative quantification and coordination not infrequently co-​occur, giving a distinctly emphatic character to the affected phrase, as in (13). (13)

Co-​occurrence of negative quantification and negative coordination [nan wihht, nan enngell, nan mann, Ne naness kinness shaffte,] no being no angel no man nor of-​no kind creature Ne mihhte þurrh himm sellfenn þa Seffne godnessess shæwenn not might through him self the seven goodnesses show ‘No being –​no angel, no man, or creature of any kind –​can by himself glimpse those seven virtues’ (CMORM,DED.L257.54)

Negative quantifiers in object position can be stranded, as in the example with ne in (14). (14)

Stranded negative quantifier as negative object felawei nauede he [noni] fellows not-​had he none ‘He did not have any friends’ (Maregrete,494.176.181)

Negative objects can also be fronted, as in the example without ne in (15).

The Loss of Negative Concord with Negative PPs (15)

281

Fronted negative object [non oðer] Ø wile ge more; none other wants she more ‘She does not desire anything else more’ (Bestiary,408.22.708.[Turtle-​Dove_​Nature])

The corpus annotation did not provide a practical way to extract all examples of negatively coordinated subjects with extraposition of the negative conjuncts. The reason is that these structures are parsed as gapping, i.e., as individual but incomplete clauses, not as subject coordination. An example of such a structure is shown in (16). (16)

Subject gapping –​not included in the dataset sunne þar ne sineð [ne mone] [ne storre] sun there not shines nor moon nor stars ‘The sun does not shine there, nor (does) the moon, nor (do) the stars’ (PoemaMorale, 232.228.279)

Negative objects include negative predicates of copula verbs because objects and predicates are annotated with the same function tag in the parsed corpora, (17). (17)

Negative predicate –​counted towards negative object I Ø am no leche I am no leech ‘I am not a physician’ (CMSIEGE,74.121)

Negative associates in existential constructions are treated as subjects. This is true for structures with overt expletives, such as there or it, (18), as well as for structures without an overt expletive, for example in cases of locative inversion, (19). (18)

Negative associate of existential construction –​counted towards negative subject a. theri nys [noon hope to escape]i. there not-​is no hope to escape ‘there is no hope of escape’ (CMCTPARS,291.C1.119) b. þeri Ø is [no wattur]i ner þe cyty þen flem Jurdane. there is no water nearer the city than river Jordan ‘There is no water nearer to the city than the river Jordan’ (CMSIEGE,82.342)

282 (19)

Richard Zimmermann Negative associate in locative inversion –​counted towards negative subject a. In þisse worlde nas [na laʒe ne na larþeu] in this world not-​was no law nor no teacher ‘In this world was neither any law nor any teacher’ (CMLAMB1,81.144) b. In Perce Ø was no wyghter man]. in Persia was no braver man ‘In Persia was no braver man’ (Alisaunder,152.3647.[Part_​1].[Chap_​15].2149)

As before, examples are not counted in multiple NC structures. The search queries yielded 2,189 examples of negative objects (842 with ne, 1,347 without ne) and 2,006 examples of negative subjects (867 with ne, 1,139 without ne). Multiple NC are structures that involve at least two negative elements (not counting ne), (5), i.e., the co-​occurrence of at least two negative phrases from the list comprising the secondary negator not, negative adverbials, negative subjects, negative objects or negative PPs. The examples in (20) provide additional illustrations of multiple NC. (20)

Additional examples of the loss of ne in multiple NC a. Naueþ [no man] [none sikerhede] not-​has no man no security ‘Nobody has any security’ (OwlNight,106.1268.702) b. And [no man] Ø myght not cope whythe Hym4 and no man might not cope with him ‘And no man could cope with him’ (CMGREGOR,219.2123) c. þu nowest [none mon] [nowitht] you not-​owe no man nothing ‘You do not owe anybody anything’ (LofsongLord,215.131.75) d. ʒe Ø holde [no parlyment] [Wiþ no christen mon] you hold no parliament with no Christian man ‘you should not hold parliament (chatter) with a Christian man’ (HowHearMass,136.286.96.[Stanza_​25])

Other kinds of negative phrases are not taken into account (e.g., negative adjective phrases). While rare, such discarded negative elements may lead to

4 Negative subject + not. The existence of sentences like (20b), NC with a negative subject and the secondary negator not, probably challenges the syntactic model proposed by Ingham (2000: 33), which seeks to rule out such structures. For more discussion, see Wallage (2017: 174).

The Loss of Negative Concord with Negative PPs

283

some recall errors and misattribution of multiple NC structures. In total, the searches found 890 hits of multiple NC (395 with ne, 495 without ne). Finally, the corpora can be queried for the structure of the greatest interest in this paper, NC with negative PPs, (6), (7). The searches are straightforward. The only complication worth mentioning concerns PPs that are not placed on the clausal level, but are more deeply embedded within another constituent. Such cases are difficult to recover in a principled way, and they are therefore not included in the counts, (21). They also do not contribute to the multiple NC data, (22). (21)

(22)

Deeply embedded negative PP –​not included in the dataset a. þer nis [NP lac [PP of met no cloþ]] there not-​is lack of meat nor cloth ‘There is no lack of food or clothing’ (CMROLLTR,9.252) b. and Ø has [NP thoghte [PP of na lufe of þe worlde]]. and have thought of no love of the world ‘and [they] have thought about not loving the world’ (Cokaygne,145.29.19) or perhaps: ‘and [they] have no thought of love towards the world’ Deeply embedded negative PP –​included in NC with negative object, not multiple NC Ne do [nan ifell dede] [PP Forr [NP lufe [PP off nan lifisshe mann]]] not do no evil deed for love of no living man ‘You should not do any evil deed for love of any living man’ (CMORM,I,177.1461)

The corpora contained 481 instances of NC with negative PPs (193 with ne, 288 without ne).

2.2. Results I will now present the results of this investigation and demonstrate that, if the decline of the negator ne is measured as described above, NC with negative PPs behaves markedly differently from the other patterns. Table 1 gives an overview over the temporal distribution of the data. It presents the number of ne occurrences out of the total number of relevant examples for the categories SN, NC and NC with negative PPs in four conventional period bins established by the Helsinki corpus.

284

Richard Zimmermann

Table 1:  Overview over the temporal distribution of the occurrence of ne in SN, NC and NC with negative PP structures Period M1 (1100–​1250) M2 (1250–​1350) M3 (1350–​1420) M4 (1420–​1500)

ne in SN 94.8 % (2,029/​2,141) 73.4 % (2,174/​2,963) 17.3 % (511/​2,949) 1.8 % (23/​1,301)

ne in NC 93.0 % (1,665/​1,790) 52.7 % (1,085/​2,057) 10.6 % (220/​2,071) 0.2 % (2/​1,048)

ne in NC with neg. PPs 69.2 % (90/​130) 46.8 % (88/​188) 11.0 % (13/​118) 4.4 % (2/​45)

These figures serve only as a rough guide for expository purposes. The statistical analysis is conducted in the form of a mixed-​effects logistic regression model on individual texts each with their own associated year estimate. Table 2 presents this model. It predicts the presence of ne from time, the construction types outlined in the previous section (with SN as the reference value) and their interaction. It also controls for correlated errors on the text level through the inclusion of random text intercepts. Table 2:  Mixed-​effects logistic regression model for the loss of ne in SN and NC Formula = Ne ~ Year * Context + (1 | Text) N=16,801 Fixed effects (Intercept) Year (standardised) Context (SN → NC with negative adverbials) (SN → NC with negative objects) (SN → NC with negative subjects) (SN → multiple negative NC) (SN → NC with negative PPs) Year: Context interaction (SN → NC with negative adverbials) (SN → NC with negative objects) (SN → NC with negative subjects) (SN → multiple negative NC) (SN → NC with negative PPs) Random effect: Texts: 159 Variance of random intercepts: 2.477

Estimate 0.30411 -​3.01751

Std. Error 0.16167 0.17251

p-​value 0.1318