Current Explorations in Middle English

List of Contributors

Part I   The Transmission of Middle English Texts

Wendy Scase

The LALME typology of scribal practice: some issues for manuscript studies¹

Abstract: The Linguistic atlas of late mediaeval English (LALME) and related work in historical dialectology is heavily intertwined with the study of Middle English literary manuscripts. LALME draws extensively on texts in literary manuscripts and interprets them using its own special model of scribal practice. The LALME typology has proved very influential in the study of Middle English literature, informing literary history, editing, textual criticism, and work in the materialities of manuscript culture. This chapter proposes that our growing knowledge of the material culture of text production and wider literate practice, however, calls into question some of the basic assumptions about scribal practice that underpin LALME. A case study of Carthusian copies of Misyn’s translations of Rolle’s Emendatio Vitae and Incendium Amoris in British Library, Additional MS 37790 illustrates the claim that manuscript studies needs, and is capable of supplying, new, subtler models of scribal practice to support a revision of LALME’s interpretations of scribal texts.

Keywords: LALME, manuscript studies, Carthusians, Angus McIntosh, literatim copying, translation copying, dialect continuum, Richard Rolle

1. Introduction

The Linguistic atlas of late mediaeval English, first published in 1986 and subsequently in revised, electronic form in 2013 (Benskin – Laing – Karaiskos – Williamson 2013–, based on McIntosh – Samuels – Benskin – Laing – Williamson 1986, henceforth LALME),² is one of the most significant publications in Middle English studies, perhaps rivalled in scope, significance, and influence only by the Middle English Dictionary (Kurath – Kuhn – Lewis (eds.) 1954–2001). An outcome of the Middle English Dialect Project initiated in 1952 at the University of Edinburgh by Angus McIntosh and Michael L. Samuels (Benskin 1981), LALME ←13 | 14→has become one of the essential resources of medieval anglophone research, its impact still growing with the appearance of related resources (notably the Linguistic atlas of early Middle English (Laing 2013), an interactive electronic version, and ongoing revisions and updates (2013 and continuing)). LALME’s achievement as a ground-breaking advance in historical dialectology and its authority as a guide for localisation of later Middle English scribal and authorial texts are widely accepted. Early, vigorous debate about LALME’s methods and data has given way to mild, even dutiful, caveats when it is cited and it continues to be widely referred to across the literature on Middle English and historical linguistics. This chapter will suggest that, despite its achievement of the status of a classic resource, we need to think again about LALME, and that the resource’s implications and uses are still far from being worked out. Notwithstanding the early debate about its methods and claims, LALME embeds assumptions that have often been repeated but still never fully evidenced or scrutinised. We need to examine those untested assumptions. We also need to think about the other uses to which the LALME data might be put. This chapter seeks to illustrate how rethinking LALME is important for the field of manuscript studies, a field hardly developed when LALME was first conceived but now one with a vigorous research agenda and large secondary literature.³ After an overview of the LALME typology of scribal practice and its impact in literary and manuscript studies, this chapter argues that some of LALME’s key assumptions are not easily reconcilable with our growing knowledge of medieval manuscript culture and literate practices. A final section illustrates this claim with a case-study of the Amherst scribe.

2. LALME: aims and achievements

The goals and assumptions of LALME will be well known to many, especially those who participated in the project, but require repetition and summary here since my intention is to disclose their less recognised implications for manuscript studies. LALME was developed as a solution to the problem of describing and understanding the history and dialects of the English language from a ←14 | 15→period from which no living speakers or audio records of speech survive and the written records of which were notoriously transmitted, often, by innumerable, often anonymous, scribes. In place of the goal of describing the phonology, morphology, syntax, and lexis of the spoken language, LALME substituted the goal of describing the written language. It developed a theory of the scribal written record as behaving like modern spoken language. Insights in the dialectology of spoken language had shown that dialects were part of a continuum rather than being varieties demarcated by hard borders. The LALME researchers interpreted the written record likewise as the product of a “dialect continuum” (Benskin 1981: xxix), viewing spellings and forms as analysable in ways similar to the ways in which, for example, the sounds of speech in modern languages could be analysed by dialectologists. In principle, the fact that the surviving manuscript texts are nearly all scribal copies and many are at one or more – perhaps many more – removes from the archetypal texts meant that most of the surviving written record appeared, on the face of it, not to be suitable for dialect analysis. Accordingly, earlier dialectologists had avoided the use of literary texts as evidence because they were viewed as being likely to be contaminated by successive copyings (e.g. Moore – Meech – Whitehall 1935). The LALME team endeavoured to solve this problem by advancing a new theory of Middle English scribal practice. According to this theory, scribes adopted one of three types of practice with regard to their exemplars. The LALME typology of scribal practice is set out in Angus McIntosh’s 1973 article “Word geography in the lexicography of mediaeval English”. Here he proposes that scribal responses to English-language exemplars fall into three categories:

McIntosh first describes these three types in his 1963 essay “A new approach to Middle English dialectology”, terming texts of type B translated and those of type C Mischsprachen (McIntosh 1989[1963]: 27–28). Margaret Laing and Michael Benskin termed McIntosh’s Type A literatim copying, or sometimes mirror copying (Benskin – Laing 1981; cf. LALME, General Introduction: 3.1.2; 3.5.1–5). The LALME typology justified the use of literary manuscripts as sources for ←15 | 16→dialect analysis. It greatly increased the number of texts thought to have value for dialect analysis. Texts of type A, the literatim copies, and type B, the translated copies, although scribal copies, could be evidence of internally-consistent written language. The literatim copy could represent consistent language simply replicated by the scribe. The translation copy could represent the language of the scribe. Laing – Williamson argued that even mixed, Type C texts could be of value for the recovery of dialects via a process of linguistic “archaeology” (Laing – Williamson, 2004, cf. Laing 1988 and see further below). Laing summed up this move as a “principle” underlying the Middle English Dialect Project and one of “two fundamental departures from traditional approaches to historical dialectology”: for the first time, “‘translated’ texts were seen to have a value equal to ‘original’ texts as source material for dialect characteristics” (Laing 1989a: ix).

Drawing heavily on literary manuscripts, the LALME project has had since its publication considerable impact in manuscript studies. The fifth York Manuscripts Conference held in 1989 celebrated the publication of LALME and an edited volume was assembled by Felicity Riddy under the rubric of “regionalism” (Riddy (ed.) 1991). Debate largely focused on the resource’s reliability as a guide to the localisation of scribal texts (e.g. Pearsall 1994: 242, Burton 1991: 179–184; cf. Benskin 1991). Localisation of texts has proved one ←16 | 17→of LALME’s main uses among literary and manuscripts scholars. Scholars have continued to rely on LALME for the localisations of their source manuscripts, albeit with some reservations, for example Nichols’ edition of “O Vernicle” (2014: 320–322) and Hanna’s study of the northern manuscripts of the Speculum Vitae (2008: 280–286). The LALME typology of literatim, translation, and mixed copying has frequently been invoked to support textual criticism and editorial practice, the implications of the typology for working out stemmata and recovering authorial spellings being discussed in LALME itself (General Introduction 4.2.1–5). The typology has proved seminal in editorial and other literary studies and only a few of the many examples can be mentioned here. Seymour (1992), editor of the Middle English translation by John Trevisa of De Proprietatibus Rerum by Bartholomaeus Anglicus, identifies a fragment of the text in Bodleian Library, MS Ashmole 1481 as a “literatim” copy of that in British Library, Addit. MS 27944. Richard Beadle invokes the notion of literatim copying in his analysis of the Canterbury Tales manuscript copied by Geoffrey Spirleng (1997). Mills (1998) gives step-by-step instructions on how an editor could use LALME in the analysis of the textual relations between witnesses. Simon Horobin invokes the typology and methods in his discussion of Chaucer’s own language and that of his scribes (2003; the LALME scribal typology is set out on 11–13) and mobilises the concept of literatim copying to propose a revised understanding of the history of copying of Piers Plowman manuscripts (2005, 2010). Daniel Wakelin invokes the typology in his discussion of scribal corrections (Wakelin 2014: 53–54).

Yet it is also in manuscript work that some issues with LALME have begun to be mentioned. Soon after the resource’s publication, Seymour leavened praise for the methods and principles of LALME with calls for more consideration of scribal practice, claiming that the “methods are readily acceptable”, while “what is said of scribes and, more widely, of scribal copying practices […] (in essence literatim, verbatim, and a mixture of the two) does leave a great deal unsaid” (1991: 74). Studies of manuscript corpora have signalled a sense of disparity between the material corpus and the linguistic analysis. Hanna (2005: 26–32) argues that Samuels’s mapping and chronological analysis of language in the London area suppresses a local literary and scribal culture of rich complexity. Hudson reports that she has developed doubts concerning the geographical localisation of the orthography of Wycliffite texts over two decades of her work on the material: “difficulties of various kinds began to emerge: the textual evidence simply did not add up to the conclusions drawn; in addition, substantial questions arose about the evidence that had been used in coming to those conclusions” (2017: 78, n. 4). Thaisen’s 2014 study explains literatim copying of the first words of Middle English verse lines in relation to issues of place-finding ←17 | 18→and avoiding eye-skip, suggesting that scribal practice is in a dynamic relation with many other factors. Drout – Chauvet propose to enhance textual criticism of Old English poetry by replacing the LALME typology with an “authority/deference” model of scribal practice (2015: 283) that allows for a dynamic relation between exemplar and scribe. Stenroos, too, urges that the contexts of text production should come into play in linguistic analysis, on the basis of “one of the major insights of modern pragmatics […] that language is always produced in context” (2016: 105): medieval literacy was not evenly distributed geographically or socially, and education and travel would have influenced written patterns. She nuances the idea of the translating scribe, suggesting that the idea of “a [scribe’s] completely consistent and unaffected usage” is “essentially an abstraction” (2016: 104–105).

What, then, are the assumptions made by LALME and related studies about the contexts of scribal activity and how well do they hold up against our current knowledge of medieval manuscript culture and literate practice? LALME assumes that scribes have their own spontaneous language: “their own usage” (General Introduction: 4.2.5); their “habitual written language” (LALME, General Introduction, Appendix I: head note); “their own spontaneous forms” (LALME, General Introduction, Appendix I: 1.4). Scribes who produce mixed language texts sometimes mirror the forms of their exemplar and sometimes translate into their own language. Translating scribes convert the language of their exemplar into their own local language, or, as we have seen, to be more precise, the language of the locality where they “learned to write” (LALME, General Introduction: 4.1.1). The large number of linguistic profiles in LALME emerges from a dialect continuum of written practice. Literatim scribes suppress their own language, only betraying it through occasional slips or “show-throughs” (Benskin – Laing 1981: 58). Benskin (1991: 26) suggests that “informal scribbles” can represent “good local language” more readily than “formal texts”. Smith (1991: 60) suggests that the scribe of Ancrene Riwle in BL, Cotton Cleopatra MS C. vi adopts the forms of his exemplar “at the expense of his ‘natural usage’”. Waldron suggests that some scribes of Trevisa’s translation of the Polychronicon were subject to some kind of control that caused them to “overrid[e] their own local habits” (1991: 87). LALME offers further hypotheses to support the argument that “any scribe who had an active command of Middle English as his habitual written language […] was more likely than not to impose his own written language on a text that he copied” (LALME, General Introduction, Appendix I: head-note). Reasons offered for why translation was likely to be imposed by a scribe include the demands of cursive writing where the unit of textual transfer was necessarily more than one letter at a time (in cursive hands ←18 | 19→letters are copied as linked units), and the growth in the number of scribes who mainly copied English, who therefore were perhaps not formed by the norms of Latin copying (LALME, General Introduction, Appendix I).

A recurrent assumption in the literature that appears to give weight to the idea that there must have been myriad local written languages is that using a local writing system, and converting texts from other dialects into a local system, had communicative advantages for scribes and readers. McIntosh suggests that popular translated texts were made “perhaps for reading aloud to a strictly local audience” (1989 [1973]: 93; cf. LALME, General Introduction, Appendix I: 1.1). Smith suggests that the scribe of the Riwle in British Library, Cotton Nero A. xiv “converts the language of his exemplar into a language that he, and presumably his readers, found easier to understand” (1991: 65). Beadle infers that the number of texts copied in the distinctive language of East Anglia were aimed at readers of the region “familiar with the spelling system” (1991: 90). However, while LALME does appear to give evidence for a large number of writing systems (writing of LALME in 1973, McIntosh claimed “we have good written evidence for well over a thousand dialectally differentiated varieties of later Middle English” (1989 [1973]: 86)), it does not support the inference that these better facilitated local communication. True, there is the occasional, much cited piece of evidence that dialect translation supported comprehension, especially the note in Cambridge University Library, MS Ii.IV.9, fol. 197v, recording that the Informacion of Richard the ermyte has been “translate oute of Northarn tunge ←19 | 20→into Sutharne that it shulde the bettir be vndirstondyn of men that be of the Selve Countre” (McIntosh 1989 [1963]: 27) and the explanation in the Cursor Mundi that this text has been “turnd […] till our aun/ Langage o northrin lede/ Þat can nan oþer englis rede” (LALME, General Introduction: 1.3.3). But, arguing against this view are the many texts that the Middle English Dialect scholars define as Mischsprachen (Benskin – Laing 1981), and the many manuscript codices that contain texts copied in a variety of dialects, either by many scribes or one (as for example British Library, Harley MS 2409, which McIntosh observes “contains four texts all in the hand of one scribe, yet each is in a self-consistent dialect and obviously therefore hardly tampered with by him” (McIntosh 1989 [1963]: 27), or the extreme case of Oxford, Bodleian Library, MS Hatton 69, in the hands of five scribes but where McIntosh and Wakelin “tentatively distinguished no fewer than nineteen different kinds of West and Central Midland English” that range from Shropshire in the west to Leicester in the east and from north Staffordshire in the north to mid Worcestershire in the south (McIntosh – Wakelin 1989 [1982]: 173–174)). Such manuscripts must have had utility for their owners and users, communicative advantages or no. The manuscript evidence suggests that there may be more to the policy announcements in the Cursor Mundi and Cambridge University Library, MS Ii.IV.9 than meets the eye.

And if they were motivated to do so, how and where would scribes have learned to write a local language and their readers have learned to read one? Scholars associated with LALME imply that training in local written language took place in school or was focused on the parish. Michael Samuels states that “[the] minute regional variations of spelling which are found earlier in Middle English […] had presumably been based on the parish” (Samuels 1981: 48; cf. Samuels 1988 [1985]: 75, “People learned the dialect spelling of the place in which they were brought up”). Smith builds on this view, referring to a “parish-based literacy”: “a process of orthographic divergence may be expected given […] a shift from the restricted literacy of the Anglo-Saxon period, focussed on a few provincial and monastic centres, to a parish-based literacy associated with the rise of characteristic twelfth-century institutions, such as grammar schools” (2000: 131). These assertions appear to be based on inferences from the interpretation of the LALME data as the record of a continuum of local written dialects. No historical evidence is offered by LALME for the teaching of writing in English in the period.⁴

The most fundamental assumption of all, underlying all of the others discussed above, is that scribal writing in the vernacular transcribes spoken language and is therefore a different kind of writing from Latin: “[c]opying Latin demands that in all essentials the spelling of the copy-text be replicated” (LALME, General Introduction, Appendix I: 1.3). LALME observes that there are “levels” of translation, with syntax and lexis perhaps remaining untranslated in a given text while morphology and spelling more often attract translation (LALME, General Introduction, Appendix I: 3).⁵ LALME infers that “[p]art of the impulse to translate must have depended on the association of particular spellings with particular forms in the spoken language” (LALME, General Introduction, Appendix I: 1.4). LALME thus assumes fundamentally different models for writing in Middle English and writing in Latin. But to what extent is this difference in model compatible with the evidence? LALME acknowledges that precise phonetic transcription was unachievable: no orthography could have possibly been refined and nuanced enough to reflect the precise characteristics of local phonology: “graphic units are not designed to carry some bits of phonic information at all […] An English-language newspaper, when it prints this or that symbol, does not thereby signal or prescribe, at the phonetic level, one and only one ‘matching’ English sound. The letter-units out of which its text is mainly made up may convey different phonic messages to different readers” (LALME, General Introduction: 1.4.3). But it is not only the practical difficulties of encoding sound in spelling that the manuscript record calls into question, it is the model of writing itself. Latin must have varied in pronunciation from one medieval community, let alone country, to another, yet the conventions for writing Latin were fairly standard with, apparently, no communicative disadvantage. And the manuscripts demonstrate that the same scribes copied Latin and English, often switching between languages on the page. How likely is it that they constantly switched back and forth between fundamentally different modes of writing?

If the data presented by LALME is not securely interpretable as the outcome of over a thousand or so local written languages developed to promote local communicative needs, how is it to be explained? This is a question for linguists, but it is also a problem that manuscript studies should tackle. My current research is concerned with developing alternative frameworks for analysis, based on study ←21 | 22→of scribal practice and graphic culture, and with proposing alternative ways of reading the manuscript corpus. In the remainder of this chapter I offer a case study of the Amherst scribe to illustrate some of the ways in which we might rethink the analytical frameworks offered by LALME and to demonstrate how that rethinking could transform our interpretation of the manuscript evidence.

The copyist whom I label the Amherst scribe produced London, British Library, Additional MS 37790, c. 1450 (LALME’s dating, based on that of the MED). Also known as the Amherst manuscript (the British Museum bought the manuscript in a sale of the goods of Baron Amherst of Hackney, d. 1909), Addit. 37790 is one of three manuscripts attributed to the scribe, the others being British Library, Egerton MS 2006 and Cambridge, St John’s College, MS 189 (G. 21). Texts in the hand of this scribe, who perhaps signs his work with the intertwined initials I. S., vary in language from one to another.⁶ The LALME entry for Addit. 37790 summarises “MS in one hand but variable language”. Richard Misyn’s translations of Richard Rolle’s Incendium Amoris and Emendatio Vitae are analysed as “S Lincs mixed with central Lincs”, the Pistill of St Bernard as “S Lincs mixed with SE Central Lincs”, the short text of the Revelations of Divine Love by Julian of Norwich as “S Lincs mixed with a more northerly element”, the Perfection of the Sons of God as “S Lincs mixed with SW Lincs” and The Mirror of Simple Souls as “S Lincs mixed with Central W Worcs”. The Visions of St Matilda (a translation of Mechthild of Hackeborn’s Book of Ghostly Grace) in Egerton 2006 (dated second or third quarter of the fifteenth century by LALME) is analysed as “S Lincs, mixed with a more southerly component”. The Pilgrimage of the Life of Man in Cambridge, St John’s College, MS 189 (dated third quarter of the fifteenth century by LALME), is described as “S Lincs, mixed with a more northerly component and a SW Midland component”. Owing (presumably) to this scribe’s incorporation of forms from various dialects, LALME does not include a Linguistic Profile for him. However, Margaret Laing attempted to sift out the scribe’s “own spontaneous usage” from the forms of his exemplars (Laing 1989b: 206), in part deriving evidence from the other two manuscript witnesses of the Misyn texts. Making a close comparison of the texts, Laing identified the exemplars of Richard Misyn’s translations of Rolle’s Emendatio Vitae and ←22 | 23→Incendium Amoris in Addit. 37790, fols. 1–18r and 18v–95r as Oxford, Corpus Christi College, MS 236, fols. 45r–56v and 1r–44v (Laing 1989b: 191–199). Laing then attempted to identify the forms of the scribe’s usage on the basis of all of the texts in Addit. 37790, placing the scribe “in the area of Grantham”, south west Lincolnshire (1989b: 208). Margaret Laing and Keith Williamson built on this study in their 2004 “archaeology” essay, analysing all of the texts attributed to the hand of the Amherst scribe. Laing and Williamson find only 29 “show-throughs” of the Amherst scribe’s preferences, most of them “confined to a single occurrence” (2004: pp. 95, 116 (table)). In their analysis, the scribe is “to some extent a literatim copyist” (2004: 92). This habit means that he produces a Mischsprache when he responds to exemplars. His “own assemblage” of features points to a language that may be fitted on the “dialect continuum” offered by LALME in south east Lincolnshire (2004: 104, 88, 98).

We have seen that the concept of a dialect continuum and of scribes who suppress to a greater or lesser extent a spontaneous local written language is problematic theoretically and philosophically from the pragmaticist’s point of view. It is also, in my view, questionable whether the “show-throughs” that Laing – Williamson list can constitute evidence for a consistent set of preferred usages. First, as noted earlier, Laing – Williamson find that the scribe’s language shows through only once in the case of many forms. Second, many of the items listed by them as the scribe’s “preferred usage” are highly variant. For example, in a list of the scribe’s claimed usages compiled from all of his known copies (Laing – Williamson 2004: 117–119), IF is spelt nine different ways (Ʒif, if, Ʒyf, yf, Ʒiff, Ʒyff, giff, gif, iff), there are seven forms for the ending of the preterite singular of weak verbs (+yd, +ed, +id, +de, +ede, +ett, +yde), HIGH is spelt four ways (hye, hie, hey, heygh), and the stem of GIVE is spelt seven ways (gyf, giff, gyffe, gyef+, gyeve, gyev+, giv+). Given that the Amherst scribe copies his exemplars very closely, disregarding their linguistic diversity, and given that his ‘own’ language is so diverse, can this scribe justifiably be said to have his own preferred usage? It is hard to find comparable prodigality among the LALME Linguistic Profiles which are purportedly those of coherent local languages. For example, LP 8640 has nine different forms for IF but only one for HIGH (GIVE and the preterite singular of weak verbs are not recorded); LP 383 has nine for GIVE but only three for IF, two for the preterite singular of weak verbs, and only one form for HIGH. And the diversity noted above just occurs in the material that remains after Laing – Williamson have sifted out the exemplar forms. Across his entire output, in addition to his ‘own’ nine forms for IF, Ʒif, if, Ʒyf, yf, Ʒiff, Ʒyff, giff, gif, iff, Laing – Williamson record another six forms deduced to be from his exemplars, gyf, yff, yif, Ʒyffe, Ʒiffe, and yiff, making a total of 15 forms for ←23 | 24→IF across the scribe’s outputs. In addition to the scribe’s four “own” forms for HIGH, hye, hie, hey, heygh, Laing – Williamson record he, hee, hegh, heghe, hyye, hy, hyƷe, heye, hiƷe, hiee and hiegh, making a total of 15 forms for the adjective across the scribe’s outputs.

From the perspective of manuscript studies, there arise further possibilities for analysis and interpretation of these and other aspects of the Amherst scribe’s practice. From the palaeographical point of view, the Amherst scribe uses two scripts in Addit. 37790. He deploys a display hand for biblical quotations, rubrics, and explicits (for example, fol. 18r) made with a wide nib, enlarged letters, and some quadrata minims. In treating biblical quotations in an enlarged script he diverges from his exemplar. For the text he uses a neat, regular Anglicana Formata. He is capable of various sizes of script. He has checked and corrected his work. For example, “for” is subpunctuated and “fro” inserted above (fol. 2v, line 8), “do” is inserted with a caret mark (fol. 5v, line 1), and dittography is corrected by crossing through (fol. 14r, bottom two lines). He copies substantial Latin texts as well as English ones (for example a compilation of extracts from Latin contemplative texts, fols. 226r–233v). He has a fine repertoire of capital letters (for example, see four examples of “To” after paraphs on folio 15v, and “This”, “That”, “Therefore”, “This” and “The” on folio 16r.) Possibly his display of graph shapes is sometimes stimulated by those in the exemplar. For example, the shapes of capital I at Addit. 37790, fol. 28r, “It is I nogh”, lines 3–4 up mirror those of the phrase in OCCC 236, fol. 6r, col. b, line 11 down. In OCCC 236 the first I graph reaches to the line above and has a bowl at the top facing to the left, with a left-facing mark on the stem further down. The second capital I extends only to x-height and has no medial mark. Addit. 37790 imitates these graph shapes and placings in relation to the line. ←24 | 25→All of these features suggest that the scribe is using visual copying and paying close attention to graph shapes.

Marleen Cré’s study of Addit. 37790 largely follows Laing’s 1989b essay in its discussion of the Amherst scribe (2006: 50–59) and repeats Laing’s interpretation ←25 | 26→of his practice without discussion. But Cré’s study of the manuscript offers some promising leads for us to develop a picture of the context of production and reception in which the scribe was working and which arguably might at least partly explain his practices. Building on work carried out by Theresa E. Halligan on Egerton 2006, Cré states, “it seems likely that the Amherst scribe was a Carthusian, and that he wrote Additional 37790, Egerton 2006 and St John’s College 189 in his cell in one of England’s Charterhouses” (2006: 53; cf. Sargent 1976: 230 who places Addit. 37790 decisively among “Carthusian products”). The Carthusians continued to copy books themselves when other religious orders were commissioning copies from scribes beyond the cloister (Cré 2006: 53), meaning that if he was a Carthusian, the Amherst scribe belonged to a particular group of practitioners who were motivated by a distinct ideology, copying texts being part of the Carthusian vocation. And if he was a Carthusian scribe, he may have been producing books for religious in his own or other houses (a Carthusian made a copy of the Speculum devotorum for a woman and Carthusian texts found their way to Bridgettine and Dominican nuns), or he may have been copying for lay audiences (Cré 2006: 53–54). We know from James Grenehalgh’s annotations and monogram that Addit. 37790 found its way into the hands of at least one Carthusian reader later in the century (Cré 2006: 291–294). Cré’s study of the reading practices of the Carthusians provides us with an insight into the kind of reader and reading that the Amherst scribe was catering for. She argues that the choice and organisation of texts in the manuscripts resources the lectio divina, a practice of reading characterised by “the process of internalizing Scripture readings and other sacred texts” (2006: 252). This involved committing text to memory for use in prayer and meditation. The texts themselves needed not always to have been available in written form for use in this way (2006: 261) and the reading process may have focused on fragments of text (253) and may have been random rather than linear (279). A fundamental principle of Carthusian reading was that the true book is Christ, who transcends text (2006: 280).⁷

Might the Carthusian context provide a context in which we can interpret and explain the Amherst scribe’s variety of scribal practice? Scribes of the Carthusian order, Malcolm Parkes suggests, display “a personal idiom” in their handwriting which is expressed in such features as “a similar range of variant letter shapes”. Parkes speculates that copying in the seclusion of a cell may have intensified the development of a personal idiom, since there is no evidence that scribes were ←26 | 27→trained by the order (Parkes 2008: 121–125).⁸ This suggests, at best, very limited support for a Carthusian explanation for the Amherst scribe’s output. Even more discouragingly, Cré sees the scribe’s “automatic and verbatim copying of his models” as at odds with the design of Addit. 37790 as a Carthusian resource, “a model and also as an aid for the reader’s spiritual ascent” (2006: 278). But not all of the scribe’s practice can be considered a personal interpretation of the Carthusian rule. The manuscripts and prints of Nicholas Love’s Mirror, also transmitted by Carthusians, exhibit similarly careful respect for their exemplars (Smith 1997: 138–139; Hellinga 1997: 155, 156, 160). Smith attributes this phenomenon to “constraint” exerted on the Love scribes by the exemplars. But this scribal behaviour challenges manuscript studies to find social and cultural explanations. What impact might being a member of this silent order have had on scribal practice and even process? The Carthusian ideology and practices of reading form a suggestive context that chimes with the Amherst scribe’s process and practices. Text destined for memorisation and internalisation in small fragments would not be experienced as a Mischsprache. The process of careful rumination and internalisation would perhaps be enhanced if the text’s variety of letter forms and orthography slowed down reading. Perhaps the repertoire of capital graphs might also have aided the internalisation of fragments, as the careful punctuation might have done. The fine variety of capital letters might also have been held to enhance the experience of close rumination on the text. And putting effort into tidying up accidentals of the text such as orthography might be considered a vain and unworthy distraction away from the true book that lay beyond text. Perhaps a focus on graph shapes was considered compatible with the rule, while a focus on encoding the sounds of speech in systematic orthography was not. Perhaps the Carthusian rule of silence encouraged a view of scribal copying as a graphic rather than a phonic art.