The Miniatures and Meters of the Old English Genesis, MS Junius 11: Volume 1: The Pictorial Organization of the Old English Genesis: The Touronian Foundations and Anglo-Saxon Adaptation. Volume 2: The Metrical Organization of the Old English Genesis: The Anglo-Saxon Foundations and Old Saxon Adaptation. [2 volumes ed.] 9783110788068, 9783110786880

Citation preview

Seiichi Suzuki The Miniatures and Meters of the Old English Genesis, MS Junius 11

Ergänzungsbände zum Reallexikon der Germanischen Altertumskunde

Herausgegeben von Sebastian Brather, Wilhelm Heizmann und Steffen Patzold

Band 138

Seiichi Suzuki The Miniatures and Meters of the Old English Genesis, MS Junius 11

Volume 1: The Pictorial Organization of the Old English Genesis: The Touronian Foundations and Anglo-Saxon Adaptation.

ISBN 978-3-11-078688-0 e-ISBN (PDF) 978-3-11-078806-8 e-ISBN (EPUB) 978-3-11-078817-4 ISSN 1866-7678 Library of Congress Control Number: 2022950671 Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available on the internet at http://dnb.dnb.de. © 2023 Walter de Gruyter GmbH, Berlin/Boston Typesetting: Dörlemann Satz GmbH & Co. KG, Lemförde Printing and binding: CPI books GmbH, Leck www.degruyter.com

For Yasuko

Preface The Old English Genesis is the sole illustrated Anglo-Saxon poem. In full appreciation of this unique concurrent execution of visualization and versification in a single manuscript, this multidisciplinary work explores the pictorial (Volume 1) and the metrical (Volume 2) organization of this illuminated verse narrative from both synchronic–structural and diachronic–comparative perspectives. While, autonomously conceived and independently implemented, each volume is solidly embedded in the respective scholarly tradition and pursues its own inherent disciplinary concerns and problematics, vigorous formal and cognitive reasoning and theorizing run commonly through both. By way of mutual corroboration and integration, the twin volumes eventually converge on the hypothesis that the earliest portion of the extant Old English Genesis (verse lines 1–966) derived from the corresponding episodes of an illustrated Touronian Old Saxon Genesis in both pictorial and metrical terms. The present volume is a formally and cognitively oriented art historical study of the Old English Genesis, with exclusive reference to the first twenty-two pictures contained in MS Junius 11 (Oxford, Bodleian Library). The primary objectives of this book are twofold. First, it will demonstrate that the first twenty-two pictures in the Old English Genesis are integrated into the system of pictorial organization that is predicated on the multidimensional, hierarchical network of opposition, complementation, parallelism, and variation in conceptual and compositional terms. Accordingly, all pictures are mutually dependent in their own different ways and with varying degrees of determinate force, with a few privileged core images located in the center of the system and controlling peripheral ones to a large extent. This systematicity of pictorial organization, it will be substantiated, is not simply a synchronically conceptualized descriptive state of affairs; rather, of no less relevance to the dynamic dimension, it would have played a formative role—largely through the intermediary of metaphor and metonymy—in motivating innovation and giving rise to novel images out of the inherited stock of resources. Thus, a subset of the Junius pictures would have been motivated on purely internal grounds without corresponding models elsewhere. In short, the pictorial system would have resided in the artist’s mind as a mental overarching emergent scheme for images in the making. Second, from a diachronic–comparative perspective, the pictorial system thus identified, as well as individual pictures and their constituents, will be shown to have been primarily founded on the Touronian tradition and to have undergone reinterpretation and elaboration in accordance with its own organization principles and preferences. Specifically, while subject to extensive reorganization and restructuring in the course of adaptation, the resources drawn on are on the whole derived from the Cotton Genesis family in general and, for the most part, from its reflections in the Touronian Bibles in particular. The resultant influence from the Tours School not only pervades the individual pictures involved but also permeates, through reorganization, all levels of visual representation, from static page layout to dynamic narrative progression, from machttps://doi.org/10.1515/9783110788068-201

VIII 

 Preface

roscopic configurations to microscopic features of individual figures, and from coordination of multiple sources to application of specific techniques of depiction. Therefore, the picture cycles under investigation in the Old English Genesis may be legitimately characterized as Touronian in their genealogical identities. As a corollary of the central claims formulated above, particularly the inference last adduced, it may follow that the original artist responsible for composing the twenty-two pictures—or at least a major subset of them—would have been a member of the School of Tours. Working at (or near) Tours in the mid-ninth century, he would have composed these illustrations as accompaniments to the relevant episodes of the Old Saxon Genesis not long after its emergence. Subsequently, the Anglo-Saxon artists would have reused these Old Saxon pictures in the earliest portion of the Old English Genesis. Of paramount importance, the conjecture presented immediately above, derived from the integrity of the twenty-two pictures both synchronically and diachronically substantiated, receives independent corroboration from the metrical organization of the earlier portion of the Old English Genesis, verse lines 1–966 (comprising Genesis A, verse lines 1–234 and 852–966, and Genesis B in its entirety, that is, verse lines 235–851). As elucidated in sections 1.1 and 12.8, drawing on the companion volume (Suzuki 2023) for fuller demonstration, the earlier part of the Old English Genesis thus delineated is, as a whole, internally homogeneous in alliterative patterning and commonly differentiated from the remainder (verse lines 967–2936) by a significantly higher incidence of double alliteration. Such a metrical bisection at the boundary between verse lines 966 and 967 may be explained in a principled way as a consequence of the narrative unity of the Old English Genesis text portion at issue, as substantiated by the derivation of the first twenty-two pictures in Junius 11 from the Touronian images originally accompanying the corresponding Old Saxon Genesis episodes. Thus, while complementary to each other in their concerns and scope of inquiry, Volume 1 on the pictorial organization of the Old English Genesis and Volume 2 on its metrical organization may be viewed as fully integrated embodying a hermeneutic circle (hopefully not a vicious circle), each supplementing and enriching the other. The 1970s saw appearance of important source studies of Junius 11, particularly Raw (1976) and Broderick (1978). Since then, however, virtually no comparably substantial work has been undertaken, although a scattering of investigations concerned with other aspects of Junius illustrations have appeared to date, notably Karkov (2001) on text–picture interrelationship and interaction as narrative strategies. While markedly different in specific claims and interpretations from earlier scholarship, this book is firmly situated in the Weitzmann–Kessler–Broderick tradition of the Cotton Genesis family research (Kessler 1977; Broderick 1978; Weitzmann/Kessler 1986). In contrast, while highly similar at first glance as far as the Touronian origins are postulated for a group of Junius pictures, Raw’s (1976) thesis is decidedly at variance with mine not only in specific claims and individual interpretations but also in general conceptual underpinnings, as outlined below in section 1.3 and detailed in subsequent chapters.

Preface 

 IX

The dearth of relevant publications in recent times does not mean that the issues relating to the origins of Junius pictures have been settled. On the contrary, recent works including Karkov (2001) have given impetus to readdressing the traditional source issues from a new perspective of conceiving the Junius pictures synchronically as a highly integrated apparatus for visual narrative rather than a contingent assemblage of individual scenes subordinate to the text, as conventionally characterized. At the same time, however, Karkov’s synchronic analyses of individual drawings and their interrelations in the Old English Genesis have left much unaddressed and unaccounted for, not least because of her not sufficiently sophisticated formal analytical techniques, particularly with respect to the paradigmatic dimension. Accordingly, drawing on my expertise in formal and structural analysis of verbal and visual representation in linguistics and archaeology (e.g., Suzuki 2008; 2014a), I will be exploring diachronic–comparative accounts within a framework of the pictorial system of the Old English Genesis that has been established with formal rigor in the synchronic part of this study. This system-based conceptualization in general, and the specific ranked relations posited between the pictures in particular, will make possible postulation of genealogical affinities hitherto unrecognized and formulation of derivational histories involved, as with the two images of God enthroned on the first two pages of the Junius manuscript—Pii (Pl. 1) and P2 (Pl. 2), the prime members of the pictorial system (section 2.1)—which are derived from the portrait of Charles the Bald in the Vivian Bible (fol. 423r, Pl. 32). Chapter 1 lays out the framework of this study by providing an overview of the major empirical data and reviewing previous scholarship focusing on Raw (1953; 1976). Chapters 2 through 11 analyze and interpret individual pictures and their interrelationship according to their thematic-based grouping. Each chapter falls into two parts: the synchronic–structural and the diachronic–comparative, with the former providing a solid empirical basis for the latter. Piecing together the cumulative body of specific analyses and arguments adduced in the preceding chapters, Chapter 12 presents synthetic views on the synchrony and diachrony of the pictorial system of the Old English Genesis, with special reference to its Touronian heritage, its enrichment of the Cotton Genesis family, and its subsequent adaptation. Chapter 13 concludes the book with some conjectures for future study. Initially supported by a grant from the Japan Society for the Promotion of Science (JSPS KAKENHI)—a Grant-in-Aid for Scientific Research C (#26370582)—and subsequently by another JSPS grant (#18K00398), this project successfully came to completion thanks to a Membership at the School of Historical Studies, the Institute for Advanced Study (IAS), that I had the honor and privilege of being awarded for the period of 2018– 2019. During my stay at Princeton, I benefited immeasurably from all conceivable logistic support and service for intensively carrying out research, the peaceful and inspiring environment conducive to creative and imaginative thinking, and, above all, the perfect academic freedom that made possible exclusive concentration on my own project for a sustained period. For all this, I would like to express my deepest gratitude to Robbert Dijkgraaf, IAS Director and Leon Levy Professor, Patrick J. Geary, Andrew W. Mellon

X 

 Preface

Professor at IAS, and all staff members at the Institute. Brett Savage deserves particular mention for his competent and professional copyediting of a book manuscript at its various stages. As always, I am deeply thankful to Yoshitaka Tanimoto, President of Kansai Gaidai University, for his understanding and generosity in granting me a leave for my tenure of Membership at IAS. Special thanks go to Herbert R. Broderick and Thomas H. Ohlgren for advice and encouragement. I am immensely indebted to the libraries, museums, and publishers that gave me permission to reproduce their copyrighted materials, as individually acknowledged in the lists and captions of Plates and Figures. Seiichi Suzuki May 2019, Princeton, New Jersey

Contents Preface 

 VII

List of Plates 

 XV

List of Tables 

 XXI

List of Figures 

 XXIII

 1 1 Introduction  1.1 The Old English Genesis, the Old Saxon Genesis, and MS Junius 11   1 1.2 The Cotton Genesis family and the Touronian Bibles   14 1.3 Raw (1953; 1976) on Carolingian sources of Junius 11 pictures   17 1.4 Notes on notations and referring expressions   19 2

God Enthroned: Pii/P2, with a brief discussion on P3.2b, P16.1, and P17.1   21 2.1 A synchronic–structural perspective   21 2.2 A diachronic–comparative perspective   37 2.3 Residual anomalies   54

3

 59 The Fall of the Rebel Angels: P3  3.1 A synchronic–structural perspective   59 3.2 A diachronic–comparative perspective   69

4

 93 The Creation: P6/P7  4.1 A synchronic–structural perspective   93 4.2 A diachronic–comparative perspective   102

5

 127 The Creation of Eve: P9  5.1 A synchronic–structural perspective   127 5.2 A diachronic–comparative perspective   130

6

 141 Adam and Eve in Paradise: P10/P11/P13  6.1 A synchronic–structural perspective   141 6.2 A diachronic–comparative perspective   147

7

 157 The Temptation of Adam: P20/P24  7.1 A synchronic–structural perspective   157 7.2 A diachronic–comparative perspective   167

XII 

 Contents

8

The Fall of Adam and Eve: P28/P31   185 8.1 A synchronic–structural perspective   185 8.2 A diachronic–comparative perspective   191

9

 201 The Covering of Adam and Eve: P34/P36/P39  9.1 A synchronic–structural perspective   201 9.2 A diachronic–comparative perspective   208

 217 10 The Judgment of Adam and Eve: P41/P44/P45.1  10.1 A synchronic–structural perspective   217 10.2 A diachronic–comparative perspective   226  237 11 The Expulsion: P45/P46  11.1 A synchronic–structural perspective   237 11.2 A diachronic–comparative perspective   241  251 12 The Touronian Foundations of the Old English Genesis: A Synthesis  12.1 A classification of primary models: the four major sources   251 12.2 The internal models (class 1) and the system of pictorial organization   253 12.3 The external models (classes 2 through 4)   266 12.4 Feeding the Cotton Genesis family: The innovation of P41/P44   276 12.5 The unused material in the Adam and Eve cycle in the Touronian Bibles   278 12.6 Auxiliary sources: The Utrecht Psalter and the Beatus Commentaries   281 12.6.1 The Utrecht Psalter   281 12.6.2 The Beatus Commentaries   286 12.7 Subsequent alterations   292 12.7.1 The scroll/scepter and the cross nimbus   292 12.7.2 The recomposition of the Temptation of Adam as the Temptation of Eve   295 12.7.3 The spindle whorl in P45.2   297 12.8 On contextualizing the production of an illustrated manuscript of the Old English Genesis: implications and conjectures   298 13 Conclusion  Bibliography 

 313  319

Contents 

 327 Index  Index of authors   327 Index of subjects   329 Index of verses   333 Plates 

 335

 XIII

List of Plates 1. 2a.

2b.

3. 4. 5. 6. 7.

8. 9. 10. 11. 12.

13. 14.

15.

16.

17.

18.

Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. ii, God Enthroned. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).   337 Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 2, God Enthroned, Facing Lucifer. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).   338 Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 2, God Enthroned, Facing Lucifer (detail). Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).   339 Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 3, Fall of the Rebel Angels. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).   340 Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 6, Creation. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).   341 Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 7, Creation. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).   342 Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 9, Creation of Eve. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).   343 Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 10, God Blesses Adam and Eve. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).   344 Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 11, God Beholds His Creations. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).   345 Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 13, Adam and Eve in Paradise. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).   346 Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 16, Fall of the Rebel Angels. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).   347 Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 17, Fall of the Rebel Angels. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).   348 Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 20, The Devil Leaves Hell/Temptation of Adam. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).   349 Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 24, Temptation of Adam. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).   350 Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 28, Adam Withstands the Temptation of the Devil/Eve Succumbs. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).   351 Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 31, Eve Tempts Adam/Adam and Eve Pray in Remorse. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).   352 Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 34, Adam and Eve Cover Themselves. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).   353 Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 36, Adam and Eve Cover Themselves/The Devil Returns to Hell. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).   354 Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 39, Adam and Eve Cover Themselves. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).   355

https://doi.org/10.1515/9783110788068-202

XVI 

 List of Plates

19. Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 41, God Calls to Adam and Eve/Judgment of the Serpent. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).   356 20. Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 44, Judgment of Adam and Eve. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).   357 21. Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 45, Expulsion. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).   358 22. Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 46, Expulsion. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).   359 23. Bamberg, Staatsbibliothek, Msc. Bibl. 1 (Bamberg Bible), fol. 7v, Genesis frontispiece (licensed under CC BY-SA 4.0).   360 24. London, British Library, Add MS 10546 (Grandval Bible), fol. 5v, Genesis frontispiece. © The British Library Board.   361 25. London, British Library, Add MS 10546 (Grandval Bible), fol. 352v, God in Majesty. © The British Library Board.   362 26. Paris, Bibliothèque Nationale, MS lat. 1 (Vivian Bible), fol. 3v, St. Jerome frontispiece. Source gallica. bnf.fr/Bibliothèque Nationale de France (licensed under CC BY 4.0).   363 27. Paris, Bibliothèque Nationale, MS lat. 1 (Vivian Bible), fol. 10v, Genesis frontispiece. Source gallica.bnf. fr/Bibliothèque Nationale de France (licensed under CC BY 4.0).   364 28. Paris, Bibliothèque Nationale, MS lat. 1 (Vivian Bible), fol. 130v (detail), God in Majesty. Source gallica. bnf.fr/Bibliothèque Nationale de France (licensed under CC BY 4.0).   365 29. Paris, Bibliothèque Nationale, MS lat. 1 (Vivian Bible), fol. 326r, Canon Table. Source gallica.bnf.fr/ Bibliothèque Nationale de France (licensed under CC BY 4.0).   366 30. Paris, Bibliothèque Nationale, MS lat. 1 (Vivian Bible), fol. 329v, God in Majesty. Source gallica.bnf.fr/ Bibliothèque Nationale de France (licensed under CC BY 4.0).   367 31. Paris, Bibliothèque Nationale, MS lat. 1 (Vivian Bible), fol. 386v, Epistles frontispiece. Source gallica. bnf.fr/Bibliothèque Nationale de France (licensed under CC BY 4.0).   368 32. Paris, Bibliothèque Nationale, MS lat. 1 (Vivian Bible), fol. 423r, Dedication frontispiece. Source gallica. bnf.fr/Bibliothèque Nationale de France (licensed under CC BY 4.0).   369 33. Rome, San Paolo fuori le Mura, Bibbia carolingia (San Paolo Bible), fol. 3v, St. Jerome frontispiece. Reproduction: Cardinali (2009: 23), © Edizioni Abbazia San Paolo.   370 34. Rome, San Paolo fuori le Mura, Bibbia carolingia (San Paolo Bible), fol. 8v, Genesis frontispiece. Reproduction: Cardinali (2009: 25), © Edizioni Abbazia San Paolo.   371 35. Rome, San Paolo fuori le Mura, Bibbia carolingia (San Paolo Bible), fol. 117r (detail), Prophets frontispiece. Reproduction: Cardinali (2009: 43), © Edizioni Abbazia San Paolo.   372 36. Rome, San Paolo fuori le Mura, Bibbia carolingia (San Paolo Bible), fol. 259v, God in Majesty. Reproduction: Cardinali (2009: 53), © Edizioni Abbazia San Paolo.   373 37. Rome, San Paolo fuori le Mura, Bibbia carolingia (San Paolo Bible), fol. 310v, Epistles frontispiece. Reproduction: Cardinali (2009: 67), © Edizioni Abbazia San Paolo.   374 38. Venice, Basilica di San Marco. Interior. Mosaics. Atrium. West Arm. South Cupola (The Creation Cupola), God Forming Adam, SM0767. Ekkehard Ritter, North Adriatic Project Fieldwork Records and Papers, 1975–1979, MS.BZ.009, Dumbarton Oaks, Trustees for Harvard University, Washington, D.C.   375 39. Venice, Basilica di San Marco. Interior. Mosaics. Atrium. West Arm. South Cupola (The Creation Cupola), Blessing of the Seventh Day, SM0770. Ekkehard Ritter, North Adriatic Project Fieldwork Records and Papers, 1975–1979, MS.BZ.009, Dumbarton Oaks, Trustees for Harvard University, Washington, D.C.   376

List of Plates 

 XVII

40. Venice, Basilica di San Marco. Interior. Mosaics. Atrium. West Arm. South Cupola (The Creation Cupola), Animation of Adam, SM0772. Ekkehard Ritter, North Adriatic Project Fieldwork Records and Papers, 1975–1979, MS.BZ.009, Dumbarton Oaks, Trustees for Harvard University, Washington, D.C.   377 41. Venice, Basilica di San Marco. Interior. Mosaics. Atrium. West Arm. South Cupola (The Creation Cupola), God Introducing Adam into Paradise, SM0775. Ekkehard Ritter, North Adriatic Project Fieldwork Records and Papers, 1975–1979, MS.BZ.009, Dumbarton Oaks, Trustees for Harvard University, Washington, D.C.   378 42. Venice, Basilica di San Marco. Interior. Mosaics. Atrium. West Arm. South Cupola (The Creation Cupola), Adam Naming the Animals, SM0781. Ekkehard Ritter, North Adriatic Project Fieldwork Records and Papers, 1975–1979, MS.BZ.009, Dumbarton Oaks, Trustees for Harvard University, Washington, D.C.   379 43. Venice, Basilica di San Marco. Interior. Mosaics. Atrium. West Arm. South Cupola (The Creation Cupola), God Creating Eve, SM0784. Ekkehard Ritter, North Adriatic Project Fieldwork Records and Papers, 1975–1979, MS.BZ.009, Dumbarton Oaks, Trustees for Harvard University, Washington, D.C.   380 44. Venice, Basilica di San Marco. Interior. Mosaics. Atrium. West Arm. South Cupola (The Creation Cupola), God Introducing Adam and Eve, SM0787. Ekkehard Ritter, North Adriatic Project Fieldwork Records and Papers, 1975–1979, MS.BZ.009, Dumbarton Oaks, Trustees for Harvard University, Washington, D.C.   381 45. Venice, Basilica di San Marco. Interior. Mosaics. Atrium. West Arm. South Cupola (The Creation Cupola), Temptation of Eve, SM0789. Ekkehard Ritter, North Adriatic Project Fieldwork Records and Papers, 1975–1979, MS.BZ.009, Dumbarton Oaks, Trustees for Harvard University, Washington, D.C.   382 46. Venice, Basilica di San Marco. Interior. Mosaics. Atrium. West Arm. South Cupola (The Creation Cupola), Eve Plucking the Fruit and Giving It to Adam, SM0790. Ekkehard Ritter, North Adriatic Project Fieldwork Records and Papers, 1975–1979, MS.BZ.009, Dumbarton Oaks, Trustees for Harvard University, Washington, D.C.   383 47. Venice, Basilica di San Marco. Interior. Mosaics. Atrium. West Arm. South Cupola (The Creation Cupola), Adam and Eve Covering Themselves with Fig Leaves, SM0791. Ekkehard Ritter, North Adriatic Project Fieldwork Records and Papers, 1975–1979, MS.BZ.009, Dumbarton Oaks, Trustees for Harvard University, Washington, D.C.   384 48. Venice, Basilica di San Marco. Interior. Mosaics. Atrium. West Arm. South Cupola (The Creation Cupola), Adam and Eve Hiding from the Presence of God, SM0792. Ekkehard Ritter, North Adriatic Project Fieldwork Records and Papers, 1975–1979, MS.BZ.009, Dumbarton Oaks, Trustees for Harvard University, Washington, D.C.   385 49. Venice, Basilica di San Marco. Interior. Mosaics. Atrium. West Arm. South Cupola (The Creation Cupola), Adam and Eve Denying Their Guilt, SM0794. Ekkehard Ritter, North Adriatic Project Fieldwork Records and Papers, 1975–1979, MS.BZ.009, Dumbarton Oaks, Trustees for Harvard University, Washington, D.C.   386 50. Venice, Basilica di San Marco. Interior. Mosaics. Atrium. West Arm. South Cupola (The Creation Cupola), Punishment of Adam and Eve, and the Curse of the Serpent, SM0796. Ekkehard Ritter, North Adriatic Project Fieldwork Records and Papers, 1975–1979, MS.BZ.009, Dumbarton Oaks, Trustees for Harvard University, Washington, D.C.   387 51. Venice, Basilica di San Marco. Interior. Mosaics. Atrium. West Arm. South Cupola (The Creation Cupola), God Expelling Adam and Eve from Paradise/Adam and Eve’s Labor, SM0800. Ekkehard Ritter, North Adriatic Project Fieldwork Records and Papers, 1975–1979, MS.BZ.009, Dumbarton Oaks, Trustees for Harvard University, Washington, D.C.   388

XVIII 

 List of Plates

52. Salerno, Museo del Duomo, Salerno ivories, The Spirit over the Waters and the Separation of Light and Darkness/Creation of the Firmament. Photo: Kunsthistorisches Institut in Florenz – Max-PlanckInstitut fle0009789x_p; Photographer: Roberto Sigismondi.   389 53. Salerno, Museo del Duomo, Salerno ivories, Creation of Plants and Trees/Creation of the Sun, Moon, and Stars. Photo: Kunsthistorisches Institut in Florenz, Max-Planck-Institut fle0009795x_p; Photographer: Roberto Sigismondi.   390 54. Budapest, Museum of Applied Arts, Salerno ivories, Creation of Birds and Fish. Inv.nr.: 18858. Photo: Ágnes Soltész-Haranghy.   391 55. New York, Metropolitan Museum of Art, Salerno ivories, Plaque with God Creating the Animals. Accession Number: 17.190.156. (licensed under CC0).   392 56. Salerno, Museo del Duomo, Salerno ivories, Creation of Eve/Temptation and Fall. Photo: Kunsthistorisches Institut in Florenz, Max-Planck-Institut fle0009801x_p; Photographer: Roberto Sigismondi.   393 57. Salerno, Museo del Duomo, Salerno ivories, Expulsion/Adam and Eve at Labor. Photo: Kunsthistorisches Institut in Florenz, Max-Planck-Institut fle0009807x_p; Photographer: Roberto Sigismondi.   394 58. Klagenfurt, Kärntner Landesarchiv, Cod. GV 6/19 (Millstatt Genesis; AT-KLA 118-A-6/19-1 St Genesis), fol. 3v, Forming of Adam. Photo: Kärntner Landesarchiv.   395 59. Klagenfurt, Kärntner Landesarchiv, Cod. GV 6/19 (Millstatt Genesis; AT-KLA 118-A-6/19-1 St Genesis), fol. 6r, Animation of Adam. Photo: Kärntner Landesarchiv.   396 60. Klagenfurt, Kärntner Landesarchiv, Cod. GV 6/19 (Millstatt Genesis; AT-KLA 118-A-6/19-1 St Genesis), fol. 8r, Adam in Paradise. Photo: Kärntner Landesarchiv.   397 61. Klagenfurt, Kärntner Landesarchiv, Cod. GV 6/19 (Millstatt Genesis; AT-KLA 118-A-6/19-1 St Genesis), fol. 9r, Naming of the Animals. Photo: Kärntner Landesarchiv.   398 62. Klagenfurt, Kärntner Landesarchiv, Cod. GV 6/19 (Millstatt Genesis; AT-KLA 118-A-6/19-1 St Genesis), fol. 9v, Creation of Eve. Photo: Kärntner Landesarchiv.   399 63. Klagenfurt, Kärntner Landesarchiv, Cod. GV 6/19 (Millstatt Genesis; AT-KLA 118-A-6/19-1 St Genesis), fol. 10r, Temptation of Eve. Photo: Kärntner Landesarchiv.   400 64. Klagenfurt, Kärntner Landesarchiv, Cod. GV 6/19 (Millstatt Genesis; AT-KLA 118-A-6/19-1 St Genesis), fol. 11r, Fall of Adam and Eve. Photo: Kärntner Landesarchiv.   401 65. Klagenfurt, Kärntner Landesarchiv, Cod. GV 6/19 (Millstatt Genesis; AT-KLA 118-A-6/19-1 St Genesis), fol. 12r, Reproval of Adam and Eve. Photo: Kärntner Landesarchiv.   402 66. Klagenfurt, Kärntner Landesarchiv, Cod. GV 6/19 (Millstatt Genesis; AT-KLA 118-A-6/19-1 St Genesis), fol. 14v, Expulsion. Photo: Kärntner Landesarchiv.   403 67. Klagenfurt, Kärntner Landesarchiv, Cod. GV 6/19 (Millstatt Genesis; AT-KLA 118-A-6/19-1 St Genesis), fol. 16v, Angel Guarding the Gate of Paradise. Photo: Kärntner Landesarchiv.   404 68. Strasbourg, Bibliothèque de la Ville, Hortus Deliciarum, fol. 17r, Adam and Eve Cycle. Reproduction: Green et al. (1979: fig. 21). Credit: Warburg Institute.   405 69. Strasbourg, Bibliothèque de la Ville, Hortus Deliciarum, fol. 17v, Adam and Eve Cycle. Reproduction: Green et al. (1979: pl. 11). Credit: Warburg Institute.   406 70. Strasbourg, Bibliothèque de la Ville, Hortus Deliciarum, fol. 27r, Adam and Eve at Labor. Reproduction: Green et al. (1979: pl. 13, no. 24). Credit: Warburg Institute.   407 71. Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 49 (detail), Cain Slaying Abel. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).   408 72. Venice, Basilica di San Marco. Interior. Mosaics. Atrium. West Arm. South Cupola (The Creation Cupola), Cain Slaying Abel, SM0817. Ekkehard Ritter, North Adriatic Project Fieldwork Records and Papers, 1975–1979, MS.BZ.009, Dumbarton Oaks, Trustees for Harvard University, Washington, D.C.   409

List of Plates 

 XIX

73. Paris, Bibliothèque Nationale, Ms. lat. 6401 (Boethius), fol. 158v, Christ in Majesty. Source gallica.bnf. fr/Bibliothèque Nationale de France (licensed under CC BY 4.0).   410 74. Utrecht, Universiteitsbibliotheek, MS Bibl. Rhenotraiectinae I Nr 32 (Utrecht Psalter), fol. 9r (detail), Psalm 17.   411 75. Utrecht, Universiteitsbibliotheek, MS Bibl. Rhenotraiectinae I Nr 32 (Utrecht Psalter), fol. 26r (detail), Psalm 44.   412 76. Utrecht, Universiteitsbibliotheek, MS Bibl. Rhenotraiectinae I Nr 32 (Utrecht Psalter), fol. 37v, Psalm 67.   413 77. Utrecht, Universiteitsbibliotheek, MS Bibl. Rhenotraiectinae I Nr 32 (Utrecht Psalter), fol. 40v (detail), Psalm 71.   414 78. Utrecht, Universiteitsbibliotheek, MS Bibl. Rhenotraiectinae I Nr 32 (Utrecht Psalter), fol. 41v (detail), Psalm 72.   415 79. Utrecht, Universiteitsbibliotheek, MS Bibl. Rhenotraiectinae I Nr 32 (Utrecht Psalter), fol. 59v (detail), Psalm 103.   416 80. Utrecht, Universiteitsbibliotheek, MS Bibl. Rhenotraiectinae I Nr 32 (Utrecht Psalter), fol. 76r (detail), Psalm 134.   417 81. Utrecht, Universiteitsbibliotheek, MS Bibl. Rhenotraiectinae I Nr 32 (Utrecht Psalter), fol. 82v, Psalm 148.   418 82. Utrecht, Universiteitsbibliotheek, MS Bibl. Rhenotraiectinae I Nr 32 (Utrecht Psalter), fol. 89v (detail), Canticum 11.   419 83. Madrid, Biblioteca Nacional de España, MS Vitrina 14-2 (Facundus Beatus), fol. 187r. Images from the collections of the Biblioteca Nacional de España (licensed under CC BY 4.0).   420 84. Paris, Bibliothèque Nationale, Ms. lat. 8878 (Saint-Sever Beatus), fol. 159r. Source gallica.bnf.fr/ Bibliothèque Nationale de France (licensed under CC BY 4.0).   421

List of Tables 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23.

Location of pictures relative to section, page/line, and verse   12 Constituency of sections 1 through 16 in terms of pages/lines, verses, pictures, and their status of preservation   14 Oppositional features of Pii and P2   26 Oppositional features of P3.2b, P16.1, and P17.1   36 The system of organization for variant pictures of God enthroned (Pii, P2, P3.2b, P16.1, and P17.1)   37 Reconstructing feature values for the common model based on their reflections in Pii and P2   39 Distribution of P3.1’s design features in the five putative models   81 Distribution of P3.2b features in Utrecht 37v and 41v   87 Correspondence table of the Creation cycle in Junius, Salerno, and San Marco   103 Distribution of feature specifications among the variant pictures of Adam and Eve covering themselves   203 Distribution of feature specifications in the three groups of pictures   207 Distribution of the Covering/Hiding, the Reproval, the Denial of Blame, and the Judgment among the major members of the Cotton Genesis family   233 Feature specifications of the Expulsion scene in the Cotton Genesis family members   243 Classification of primary sources for Junius pictures   251 The major sources of the first twenty-two pictures (including their constituent pictures) in Junius 11   252 The pictorial organization of the Fall of the Rebel Angels cycle with respect to representations of God and related (extraterrestrial) figures   256 The pictorial organization of the Adam and Eve cycle with respect to representations of God and related (extraterrestrial) figures   261 The pictorial organization of the Creation cycle with respect to representations of God   262 The incipient system of pictorial organization in the Touronian Bibles   264 Distribution of figures of God in the Adam and Eve cycle, according to their horizontal locality/ directionality and in terms of token frequencies   274 Parameter estimates report for the regression coefficient on S in Beowulf, the Heliand, Genesis A, and Genesis B   305 Frequency of double alliteration per 100 a-verses according to sections in the Old English Genesis   306 Parameter estimates report for the regression coefficient on S in Genesis A1/Genesis B, Genesis A2, and Genesis III   310

https://doi.org/10.1515/9783110788068-203

List of Figures 1.

2.

3. 4.

5.

6. 7.

8.

9.

10.

11. 12. 13. 14. 15. 16. 17.

Figural and compositional correspondences between Vivian 423r (adapted, left) and Junius P2 (detail, right). Paris, Bibliothèque Nationale, MS lat. 1 (Vivian Bible), fol. 423r (detail, adapted), Dedication frontispiece. Source gallica.bnf.fr/Bibliothèque Nationale de France (licensed under CC BY 4.0). Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 2, God Enthroned, Facing Lucifer (detail). Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).   42 Junius P2 (detail), reconstructed (left) and actual (right). Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 2, God Enthroned, Facing Lucifer (detail, adapted). Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).   55 Mirror image of Bamberg 7v2 (left half). Bamberg, Staatsbibliothek, Msc. Bibl. 1 (Bamberg Bible), fol. 7v (detail, adapted), Genesis frontispiece (licensed under CC BY-SA 4.0).   138 Approximation to the original image in the Cotton Genesis, fol. 9v, based on Grandval 5v (above) and the Hortus Deliciarum 17r and 17v (below). © The British Library Board: Add MS 10546 (Grandval Bible), fol. 5v (detail, adapted). Reproduction (detail, adapted): Green et al. (1979: fig. 21 and pl. 11). Credit: Warburg Institute.   172 Alternative approximation to the original image in the Cotton Genesis, fol. 9v, based on Grandval 5v (above) and the Hortus Deliciarum 17r and 17v (below). © The British Library Board: Add MS 10546 (Grandval Bible), fol. 5v (detail, adapted). Reproduction (detail, adapted): Green et al. (1979: fig. 21 and pl. 11). Credit: Warburg Institute.   173 Grandval 5v (detail), Admonition (left) and Temptation of Eve (right). © The British Library Board: Add MS 10546 (Grandval Bible), fol. 5v (detail, adapted).   176 Recomposition of the Cotton Genesis Admonition and Temptation, prior to Junius and San Marco. Reconstruction based on Grandval 5v. © The British Library Board: Add MS 10546 (Grandval Bible), fol. 5v (detail, adapted).   179 Recomposition of the Cotton Genesis Admonition and Temptation, immediately preceding Junius. Reconstruction based on Grandval 5v. © The British Library Board: Add MS 10546 (Grandval Bible), fol. 5v (detail, adapted).   180 Compositional parallelism between P24 (left, horizontally reversed) and P31.1 (right). Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 24 (detail, adapted), Temptation of Adam; and p. 31 (detail), Eve Tempts Adam. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).   189 Lucifer in P3.1 (middle) and his messenger in P31.2 (left, horizontally reversed) and P28 (right). Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 3 (detail), Fall of the Rebel Angels; p. 28 (detail), Temptation of Eve; and p. 31 (detail, adapted), Adam and Eve Pray in Remorse. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).   190 Bivariate analysis of sections and double alliteration in Beowulf   302 Bivariate analysis of sections and double alliteration in the Heliand   303 Bivariate analysis of sections and double alliteration in Genesis B   303 Bivariate analysis of sections and double alliteration in Genesis A   304 Bivariate analysis of sections and double alliteration in Genesis A1 and Genesis B combined (sections 1–16)   308 Bivariate analysis of sections and double alliteration in Genesis A2 (sections 17–41)   309 Bivariate analysis of sections and double alliteration in Genesis III (sections 15–41)   309

https://doi.org/10.1515/9783110788068-204

1 Introduction 1.1 The Old English Genesis, the Old Saxon Genesis, and MS Junius 11 This preliminary chapter introduces the Old English Genesis, the Old Saxon Genesis, MS Junius 11, the Cotton Genesis, and the Touronian Bibles, along with some related key notions, provides a rationale for defining the problematics to be explored in this book— particularly the delimitation of the first twenty-two pictures from the remainder—and places the major proposals in perspective by differentiating them from the apparently similar but fundamentally distinct thesis of Raw (1953; 1976). The Old English Genesis (Krapp 1931; Doane 1991; 2013)—often referred to in earlier literature as the Caedmonian Genesis—is an alliterative poem composed in conformity with the principles of Old Germanic versification (Suzuki 2014a: 9–15; 2014b). It is by no means a simple translation or paraphrase of the Genesis text, but an original work of vernacular poetry to be appreciated on its own terms. It contains 2936 verse lines, the longest Old English poem after Beowulf (3182 lines; Fulk/Bjork/Niles 2008). It begins with the Fall of Angels and ends in the Returning of Isaac to Abraham, thus covering the Book of Genesis from 1:1 to 22:13. Accordingly, the Old English Genesis falls short by half of the whole of biblical Genesis ending in 50:26, but is augmented with extensive additions and elaborations, particularly in the beginning of the poem concerning the Fall of the Rebel Angels, and the Temptation and Fall of Adam and Eve. The Old English Genesis is contained in an illustrated manuscript, Oxford, Bodleian Library, MS Junius 11 (Gollancz 1927; Krapp 1931; Ker 1957: 406–408; Temple 1976: 76–78; Lucas 1980; 1981; Raw 1984; Karkov 2001: 19–32; Lockett 2002; Doane 2013: 1–41; Gneuss/Lapidge 2014: 491–493; Bodleian Library MS. Junius 11). The manuscript is dated to around the year 1000; and it comprises 229 pages. The Junius manuscript is thus a Late Anglo-Saxon product, written predominantly in Late West Saxon, with some admixture of Early West Saxon and Anglian. In addition to Genesis, Junius 11 contains three other biblical poems—Exodus, Daniel, and Christ and Satan—which are arranged after Genesis in that order. Following the frontispiece (p. ii), Genesis occupies pages 1 through 142 of the 229 pages. The first three poems were written by the same scribe, while the last piece, Christ and Satan, was copied by three others. Given that the last poem is referred to as Liber II at the end of the text, the preceding three are generally subsumed under Liber I by extrapolation from and as opposed to Liber II, actually inscribed in the manuscript. There are altogether forty-eight illustrations, some being full-page miniatures (for exemplification, see below), others half-page pictures. Although numerous spaces for illustrations are provided on pages subsequent to page 88, which hosts the last picture (i.e., the forty-eighth one, Abraham Approaching Egypt), they are left blank (e.g., pp. 99, 101, 102, 103, etc.): there are forty such blank spaces in Genesis, thirteen in Exodus, and https://doi.org/10.1515/9783110788068-001

2 

 1 Introduction

thirty-one in Daniel (Karkov 2001: 203–206). All of the completed pictures—forty-eight in number—belong to Genesis, and the other three poems in the same manuscript are totally unillustrated despite space allocations for illustration in Exodus and Daniel (no space is provided in the last poem, Christ and Satan), as noted above. As for actual illustration, two illuminators were involved: the first one worked up to page 68, taking care of the first thirty-eight pictures; the second hand was responsible for the remaining ten. Being a piece of Anglo-Saxon poetry and accompanied by pictures, the Old English Genesis is the only extant illustrated Old English poem.1 In terms of composition, the poem is divided into two parts, Genesis A (verse lines 1–234, 852–2936; pp. 1–12/18, 40/8–142/8; key: page/line in the manuscript) and Genesis B (verse lines 235–851; pp.  13–40/8; Sievers 1875: 6–7). Genesis A is a native (i.e., Old English) original, presumably produced in the eighth century (Doane 2013: 51–55). This predecessor of Genesis A may well be called the Anglian Genesis (section 1.1, Volume 2). In contrast, Genesis B is an Old English translation of the Old Saxon biblical alliterative verse, composed on the continent in the mid-ninth century (ca. 830s; Sievers 1875; Doane 1991; see also below). Thus, the extant Old English Genesis resulted from an interpolation of the translated material. Yet the resultant work apparently would have been intended to be a unitary one for all the obvious differences involved between the two constituents—as demonstrated by Sievers (1875: 7–15)—in matters of style, meter, diction, and most importantly the ways in which the biblical materials are used for story (re)telling. The Old Saxon Genesis is attested independently, although in fragments amounting in total to 337 verse lines in the edited text (Doane 1991: 232–252; Behaghel/Taeger 1996), shorter than Genesis B in total verse length. In a Latin manuscript containing computus texts (Vatican, Biblioteca Apostolica, Palatinus Latinus 1447, fols. 1r, 2r, 2v, 10v), three passage fragments taken from an Old Saxon Genesis—Adam and Eve (see below), Cain and Abel, and Abraham and Sodom—are added to blank spaces of the manuscript, of which verses 1a through 26a (in terms of the edited text) correspond to 790a–817a in the Old English Genesis (= Genesis B). For details, see Doane (1991: 13–28) and section 3.17.1, Volume 2. It is assumed for now, and eventually made most plausible, that the Old Saxon Genesis episodes—the Fall of the Rebel Angels, the Creation, and Adam and Eve—would have been accessible at Tours in fuller form for illustration. There are no indications at all in the manuscript that the three portions—Genesis Aa (verse lines 1–234), Genesis B (verse lines 235–851), and Genesis Ab (verse lines 852– 2936)—are differentiated formally, despite the sophisticated use for other purposes that the decorated initials and various punctuation marks are made to serve in the manuscript, most importantly for signaling the beginning of a section, the unit of poetic organization next above the line and the highest next below the whole poem (for details, see further below). Simply put, at both interfaces—Genesis Aa/Genesis B and Genesis

1 Of course, there are other illustrated Old English texts, but they are all prose, such as Ælfric’s Old English Hexateuch (London, British Library, Cotton MS Claudius B IV).

1.1 The Old English Genesis, the Old Saxon Genesis, and MS Junius 11 

 3

B/Genesis Ab—Genesis A and Genesis B are assigned to the same sections. Nor is there any explicit marking for a transition from one version of Genesis to the other within the same section. To be more precise, while Genesis Aa changes to Genesis B between verse lines 234 and 235 in the edited text, there are codicological and literary grounds for believing that two leaves were lost after Genesis A verse 234b (p. 12/18) and before Genesis B verse 235a (p. 13/1), and that Genesis B actually would have started somewhere on one of the lost pages (Doane 2013: 11). The end of verse 234b (Genesis A) on page 12 corresponds to that of Genesis 2:14, while the extant initial part of Genesis B (235a–236b) approximately relates to God’s warning in Genesis 2:16–17 (Doane 1991: 255). Since verse 234b (Genesis A) on page 12 in all likelihood—as made plausible by the incompletion of the line involved, leaving much empty line space after the end of verse 234—would have constituted the ending of section 4 (Doane 2013: 11),2 and the current beginning verse of Genesis B (235a) is located in section 5, the four lost pages in their entirety—corresponding to Genesis 2:15 and whatever elaborations added to it—would have been allotted to section 5 as well. Therefore, the end of Genesis Aa (lost) and the start of Genesis B (lost) may have been positioned in the same section, namely, section 5. On fuller reflection, however, this may prove to be a rather naive inference. Given that the lost leaves most likely would have centered on the biblical text of Genesis 2:15 (Doane 2013: 9), they would have been expanded with extensive elaborations characteristic of Genesis B. In this light, even taking into account picture spaces, Genesis B presumably would have occupied a substantial part of the lost leaves. It may even seem conceivable that Genesis B started at the head of the lost leaves. In that event, the juxtaposition of Genesis A and B in the same section would not have taken place in the transition from Genesis Aa to Genesis B. In any event, in the absence of direct attestation in the extant manuscript, exactly how the interface would have been treated is utterly unrecoverable. In contrast, attestation to the other transition, that from Genesis B (verse 851b) to Genesis Ab (verse 852a), is on record: it occurs in the middle of a line, the eighth line on page 40 (i.e., p. 40/8). Continuing in ways indistinguishable from any manuscript-line– internal juncture within the same text, Genesis B connects with Genesis A without any hint of a textual disruption (Doane 2013: 157). Indeed, if there had been any need to make a formal distinction between Genesis B and Genesis A, the scribe easily could have inserted the current picture on page 39 precisely at the transition from Genesis B to Genesis Ab by transferring it from after verse 841b (p. 39/11) to after verse 851b (p. 40/8). The resulting layout of page 39 would have been comparable to the current page 12, leaving relatively small space at the bottom of the page. That the scribe actually did not put such readily available tactics into practice—he had executed analogously on

2 In addition, as Broderick (1978: 419–422) argues, the blank space below the text on page 12 may be interpreted as an unfilled picture space.

4 

 1 Introduction

page 12—and ended up masking in effect the textual discontinuity involved may well be ascribed to his design to fully integrate the text. Further on the redaction concomitant with the translation and interpolation of the Old Saxon material, it would not necessarily be the case that Genesis B in its surviving form reflected the original integrity faithfully. Rather, it may be conceivable that diverse, discrete parts were selected from the original for unification, translation, and interpolation to best suit the narrative development of Genesis A. Of particular interest in this regard is the overlap in content between Genesis A 1–77 and Genesis B 246–337, both of which tell of the Fall of the Rebel Angels. One may conjecture that in order to make the following episode of Satan’s lament and revenge most intelligible and coherent, the initial story of the fall of Lucifer and his cohorts, comparable to that in Genesis A, was reproduced as contextual background by inserting it between God’s prohibition to Adam and Eve (ending at Genesis B 245) and Satan’s lament (beginning at Genesis B 338). Hosting the only extant Anglo-Saxon illustrated verse, Junius 11 is a unique manuscript that is highly sensitive to and sophisticated in both metrical and pictorial structuration. It is far from true that a picture may be inserted anywhere on the page regardless of the metrical properties of the adjacent verses. On the contrary, strict restrictions are imposed on where a picture may appear on the page in relation to the verse text. Some deviations from the normal patterns may arise exceptionally, but for good reason, and these aberrant cases in turn may shed light on the compositional basis of the poem, as will be argued below. Excepting thirteen full-page illustrations (Pii, P3, P7, P20, P31, P41, P44, P51, P53, P54, P61, P87, P88)3 and one page in which the three-lined body of the text overlaps the middle of the otherwise full-page picture (P77), one may distinguish two classes of pictures based on to which vertical end, the top or the bottom of the page, they are placed closer. Two pictures (P63 and P74) are aligned with the top of the page and four others (P58, P59, P60, and P62) are set close to the top, whereas the vast majority (twenty-seven examples: P2, P6, P9, P10, P11, P13, P16, P17, P24, P28, P34, P36, P39, P45, P46, P47, P49, P56, P57, P65, P66, P68, P73, P76, P81, P82, and P84) are aligned with the bottom of the page, and one additional instance (P78) is arranged toward it. Thus, to generalize, a picture is aligned with the bottom, as opposed to the top or the middle, of the page (Henderson 1975: 118; Broderick 1978: 413). In other words, a picture may be preceded but not followed by text on the same page. To formulate in more precise terms by taking into account the variation depending on specific localities in the manuscript as treated below, a picture is far more likely to be aligned with the bottom than the top of the page; put another way, it displays a conspicuous tendency to be placed at the lower part of the page than at the upper. A verse line, as printed in the edited text, consists of two verses, the a-verse (on-verse) and the b-verse (off-verse). This organization of verse and line conditions

3 On the notations such as Pii and P3, see section 1.4 below.

1.1 The Old English Genesis, the Old Saxon Genesis, and MS Junius 11 

 5

picture arrangement. A picture is normally put after the b-verse and before the a-verse, as opposed to the line-internal juncture between the a-verse and the b-verse. Only exceptionally does a picture appear after a-verse and before the b-verse, as shown below. In other words, a picture may not disrupt the integrity of a metrical line by intervening between its constituents, the a-verse and the b-verse. Thus, picture insertion takes place in full harmony with line integrity. As empirical support for this generalization, one may point out the following distribution that obtains between the location of pictures and verse boundaries. Of the forty-two pictures in singles or occasionally in doubles—as two pictures may concatenate in succession—in the illustrated portion of the Old English Genesis (up to P88; see above), thirty-seven, that is, over 88 %, appear before an a-verse (1a, 49a, 135a, 186a, 206a, 210a, 246a, 325a, 389a, 491a, 599a, 765a, 842a, 872a, 939a, 952a, 972a, 994a, 1036a, 1063a, 1101a, 1143a, 1159a, 1167a, 1179a, 1197a, 1215a, 1225a, 1237a, 1314a, 1327a, 1363a, 1497a, 1602a, 1697a, 1719a, 1767a; for details, see Table 1), four before a b-verse (309b, 663b, 731b, 1588b), and the remaining picture is located in the middle of an a-verse (1830a; þæt me wrāðra [P87/P88 are inserted here] sum).4,5 The metrical organization interacts with pictorial arrangement not only in broad terms of verse distinction between the a- and the b-verses, but also at a finer level of verse constituency. More specifically, a verse constitutes a configuration of metrically strong positions (lifts, occupied by a stressed syllable) and weak positions (drops, occupied by an unstressed syllable or a sequence thereof) in a highly restricted way (section 1.2, Volume 2). Of immediate concern here is whether a verse begins with a lift or a drop (including anacrusis, extra unstressed material at the beginning of a verse). This binary distinction in verse opening has a significant bearing on the placement of pictures. Specifically, one may distinguish two classes of a-verse that appear immediately after pictures, as treated above. Of the total of thirty-seven such a-verses, twenty-five instances begin with a drop, and the remaining twelve with a lift. By contrast, the a-verse appearing elsewhere in the illustrated portion of Genesis—that is, up to 1829b—displays a markedly contrastive preference for lift initiality, with 1023 examples beginning with a lift as opposed to 765 drop-initial instances.6 Thus, an a-verse following a picture starts with a drop twice as frequently as with a lift. By contrast, the a-verse appearing elsewhere exhibits a contrary pattern of more frequently having a lift in initial position. To generalize, the a-verse is more likely to begin with a drop when standing immediately after a picture than when it appears elsewhere (p < 0.01). Brought together, the two conditions on the arrangement of pictures relative to the verse text may dictate that

4 In addition, there is a single case (P77) in which a block of text accounting for three lines intrudes into the middle of a picture. Since this picture immediately follows P76, the two are treated in combination, and the intervening lines in question are ignored. 5 Although one leaf was lost after page 10 (Doane 2013: 11), the occurrence of a b-verse (205b) immediately before P10 may warrant positing an a-verse after the picture in the lost original, as in the current manuscript. 6 One should be reminded here that verse 1830a is split by the intervening pair of pictures, P87 and P88.

6 

 1 Introduction

pictures are most likely to be placed before an a-verse that begins with unstressed linguistic material. The demarcating role that an a-verse beginning with an unstressed syllable, or a sequence thereof, fulfills in picture placement recurs at another level of organization in the Old English Genesis, that is, the division of the verse text into sections. The Genesis text is divided into sections, the highest unit of organization in the metrical hierarchy (section 2.16, Volume  2). The sections are normally marked by zoomorphic or plain capitals at the beginning and by punctuation at the end—which usually takes place in the middle of a line, and is accordingly followed by the remaining line left blank (for details on the ways the section boundaries are treated in the manuscript, see Doane 2013: 11–18). The sectional divisions are counted continuously throughout Liber  I of Junius 11—the first part of the manuscript comprising Genesis, Exodus, and Daniel, as noted above—and often explicitly numbered, particularly toward the end of the text (see Tables 1 and 2). Based on the minor variability among sections in terms of their constituent line numbers and the many instances of section division executed in defiance of narrative logic, Doane (2013: 21) concludes that sections were intended as “units of length for fixed periods of reading (lectiones),” rather than “episodes within a narrative (sententias).” Thus, the sections may be viewed primarily as a metrical or prosodic level of organization. Such occasional mismatches between sectional divisions and narrative development, and the fairly uniform size of sections ranging between 70 and 85 lines for the most part (Doane 2013: 18), are comparable to Beowulf (Doane 1991: 38): in it, sections usually fall between 60 and 90 lines (Fulk/Bjork/Niles 2008: xxxiv) and are often divided at the cost of narrative continuity (Fulk/Bjork/Niles 2008: xxxv). As suggested in the preceding paragraph, the section may be characterized as the highest unit in the metrical hierarchy. Actually, there is evidence that the division is at least partly predicated on metrical grounds. At issue is the now familiar dichotomy of verses into the drop-initial versus lift-initial ones. The proportion of drop-initial to lift-initial verses differs significantly between section-initial a-verses and other a-verses: there are 29 drop-initial and 8 lift-initial a-verses in section-initial position, whereas there are 1236 and 1661 respective instances elsewhere.7 The distribution differs significantly with a p-value of less than 0.001. Thus, the same generalization based on the drop-initiality of the a-verse underlies picture placement and section division: much as pictures are likely to be placed before the a-verse beginning with an unstressed syllable, so are section boundaries. In other words, the verse-initial weak position of the a-verse serves as a marker of a break at a level higher than the verse line. As observed above, the Old English Genesis fixed in its present form in the manuscript would have been intended as an integrated unitary work of poetry, with the

7 The counting is based on the whole corpus of the Old English Genesis including hypermetric and other irregular verses. For carefully controlled examinations, see section 2.16, Volume 2. The respective results obtained (p < 0.001), however, make no difference.

1.1 The Old English Genesis, the Old Saxon Genesis, and MS Junius 11 

 7

internal break between the two heterogeneous constituents, Genesis A and Genesis B, having been smoothed out synchronically in codicological and textual terms. This, however, would by no means entail that this work was comparably homogeneous in its (re)making seen from a diachronic perspective. On the contrary, there is substantive evidence, to be adduced below, that the first twenty-two pictures would have constituted a closed group, which, detached from the remainder, is subject to its own organization and reorganization, and integrated by a common derivational history. While a whole justification for the exclusive treatment of the twenty-two pictures ultimately depends on credibility and acceptability of the major thrust of this book, a preliminary motivation may be required at this stage to give initial plausibility to the ensuing enterprise. Of the first twenty-two pictures, from Pii through P46, all but the frontispiece (Pii) may be subsumed under one of the three picture cycles, the Fall of the Rebel Angels cycle, the Creation cycle, or the Adam and Eve cycle. While alternative conceptualizations may be feasible, the postulation of these three cycles will do for the sake of exposition.8 In this light, the closing picture of the Adam and Eve cycle assumes utmost significance for delineation of the entire picture series. Obviously, it is the Expulsion (P46) that marks the closing of the Adam and Eve cycle, as it puts an end to their life in paradise, and P47 (Birth of Abel) represents the couple’s new life in the outside world with the next generation of their offspring. Accordingly, it is conceptually natural and legitimate to make a categorial division between P46 and P47 in terms of pictorial narrative. Needless to say, other narrative breaks are conceivable. Further justification is necessary to privilege this boundary over others imaginable located in either direction, forward or backward. Given the initial plausibility of a demarcation between P46 and P47, one may invoke two complementary pieces of substantive evidence for delineating the initial picture group at P46, one metrical, the other pictorial. As may be recalled, there is a marked tendency to align a picture with the bottom of a page, but this generalization requires qualification because of the varying conformity to the bottom-alignment contingent on the location of pictures in the whole manuscript, as suggested above. Specifically, it may be of interest to examine the extent to which non-full-page pictures, that is, pictures that appear with the text (of whatever length) on the same page, deviate from alignment with the page bottom, according to the bipartition of the pictures into the earlier (Pii through P46) and later (P47 through P88) groups. There are fifteen (P2, P6, P9, P10, P11, P13, P16, P17, P24, P28, P34, P36, P39, P45, and P46) and nineteen (P47, P49, P56, P57, P58, P59, P60, P62, P63, P65, P66, P68, P73, P74, P76, P78, P81, P82, and P84) non-full-page pictures in the first and second groups, respectively. While not a single example is attested 8 For example, the second and third cycles, or even the first as well, may be generalized as a single overarching one of the Creation. Furthermore, the status of P2 is ambivalent, partly in association with Pii and partly in grouping with P3, but nothing crucial depends on the selection of particular classificatory schemes of pictures. Of vital importance is the overall integrity of the twenty-two pictures involved— irrespective of their affinity with particular picture cycles—in metrical and pictorial terms, as substantiated below.

8 

 1 Introduction

to that is followed by the text in the first group, one finds seven deviations from the norm in the second group: P58, P59, P60, P62, P63, P74, and P78. The distribution proves to be of significant difference on statistical grounds (p = 0.0106; Fisher test). Thus, the earlier group manifests the alignment with the bottom of the page more consistently. Still more interesting is to compare the first group with another one ranging from P47 through P68, the remaining set of pictures that were illustrated by the same first illuminator, as noted above. In other words, one may be interested to see whether the strict observance of the rule in the first group can still be recognized as distinct from the subsequent remainder executed by the same hand. By excluding from consideration the pictures illustrated by the second illuminator (P73 through P88), one may safely remove the otherwise possibly distorting factor of stylistic difference between individuals. As it turns out, with a p-value of less than 0.01, it is justified to claim that, compared with the subsequent group, the earlier one exhibits a significantly greater likelihood of aligning a picture with the bottom of a page in conformity with the rule, even under the optimal condition of making a comparison within the same artist. Thus, the group of pictures from Pii through P46 stands out conspicuously by its maximal adherence to disallowing that a picture be followed by text on the same page. The lower part of the page is privileged as an optimal site for a picture in the earliest series of pictures up to P46, in distinction from the subsequent remainder.9 The strictest prohibition to arranging text below a picture is most tellingly testified to by page 31. On this page, some space is left blank under the picture, the extra space that readily could have accommodated four or five lines of the running text. That this actually was not filled in may be interpreted as a consequence of the strict requirement for aligning a picture with the bottom of the page. It may be of interest in this connection to compare with P11, P34, and P36, which constitute the contrary pattern of a few lines being allowed to occur above the picture. Thus, there is a demonstrable complementarity between top and bottom of the page with respect to text/picture alignment. Moving onto the second evidence, metrical in nature, one must begin by introducing into discussion some basics of Old English meter (e.g., Suzuki 2014b; sections 1.2–1.4, Volume 2). As may be recalled, the verse line consists of two verses, the a-verse and the b-verse. Normally, each verse contains two lifts, the metrically strong positions realized by stressed syllables. To form a metrical line, the two constituent verses must alliterate by sharing the initial consonant of the lifts. While the first lift of the b-verse invariably and solely must be involved in alliteration to the exclusion of the second lift, the a-verse has at its disposal two alternative patterns of alliteration, whereby either both lifts (called double alliteration) or just one of them (called single alliteration) may participate in alliteration. Of concern in the present context is the varying proportion of single and double alliteration in the a-verse between different portions of Genesis

9 The strict alignment of a picture with the bottom of a page has interesting implications for exploring page arrangement and narrative sequencing of P6/P7. For details, see section 4.2.

1.1 The Old English Genesis, the Old Saxon Genesis, and MS Junius 11 

 9

A. Since, as commonly known, Genesis B is sharply distinguished from Genesis A by its marked predilection for double alliteration, one must separate it out from the remaining part of the Old English Genesis, namely, Genesis A, on which one must now concentrate. Of paramount interest is to see whether the verse text covered by the pictorial narrative up to P46 is significantly distinguished in terms of alliterative pattern from the subsequent remainder. While there may be some room for disagreement in detail,10 it seems reasonable on the whole to delineate the text between 966b and 967a. That is, the text portions comprising verse lines 1–234 and 852–966 put together—designated in the present discussion as Genesis A1—are aligned with the earlier picture group (Pii through P46). In this first division of Genesis A (i.e., Genesis A1), there are 167 and 164 a-verses with single and double alliteration, respectively. As far as the remaining component of Genesis A, ranging from verse 967a to the end—designated as Genesis A2—is concerned, single and double alliteration are realized in 1122 and 587 a-verses, respectively. The distribution patterns between the two groups differ with a statistical significance of a p-value of less than 0.001. More specifically, double alliteration is more likely to occur in Genesis A1 than Genesis A2. Therefore, the bipartition on pictorial grounds is corroborated persuasively in metrical terms, thereby mutually reinforcing the division, and in conjunction providing a multifaceted empirical basis for treating the first twenty-two pictures in their own right, as will be explored assuredly throughout this book. At this point, it may be revealing to examine how Genesis B differs from Genesis A1 and A2 along the metrical parameter at issue. One finds in Genesis B the total numbers of 237 and 305 a-verses with single and double alliteration, respectively. As expected, this translation-based material displays a significantly greater frequency of double alliteration as a transparent reflection of the original Old Saxon (Suzuki 2004), even higher than Genesis A1 at a confidence level of 90 % (p = 0.059), and let alone Genesis A2 (p < 0.001). As it turns out, the portions of Genesis A that are adjacent to Genesis B (i.e., Genesis A1) exhibit a stronger association with double alliteration than the rest that is further removed from Genesis B (i.e., Genesis A2). This positive correlation between proximity to Genesis B and similarity—rather than dissimilarity—in alliterative pattern to Genesis B would be by no means accidental. The clustering of like features may lead one to hold assimilation responsible for the convergence. More specifically, the bare difference between Genesis A1 and Genesis B in alliterative pattern (p = 0.059) would likely have been a result of converging metrical practice on both parts. In anticipation of a suggestion that will be made in Chapter 12 (for details, see sections 4.2–4.4, Volume 2), this partial resemblance of Genesis A1 to Genesis B in distinction from Genesis A2 may well be attributed to an editing and reworking executed concomitant with and subsequent to the interpolation of Genesis B into Genesis A. It seems natural that the editing presumably aimed at better integration would have been done more meticulously and consistently in the portion neighboring the interpolated extraneous material.

10 For example, Klaeber (1913: 27, 31) put a narrative break at 964b.

10 

 1 Introduction

Most importantly, the delineation resulting from maximal editing was motivated on the common conceptual basis of the narrative content, verbal and visual, namely, the illustrated Old Saxon Genesis episodes as source material. The adjacency to the interpolated text in itself, however, cannot afford to specify the scope of intensive revision in precise and principled terms. The grouping of Genesis A1 and Genesis B in opposition to Genesis A2, claimed to be founded empirically, is reinforced further by supplementary evidence. The foregoing examination brought to light the correlation between picture placement and the following a-verse by identifying its drop initiality as a favorable condition for promoting picture insertion. In the present concern with the metrical parallelism between Genesis A1 and Genesis B to the exclusion of Genesis A2, one may consider anew the issue of drop initiality from a different perspective. At stake is the length of the verse-initial drop, in terms of the number of syllables aligned with it. As noted above, the drop may consist of one or more unstressed syllables. Without entering into technical details, there are four different metrical configurations (verse types and their variants) involved in drop initiality (/ = lift; × = drop; for expository purposes, a distinction between normal [×] and heavy [\] drop is ignored; for details, see Suzuki 2014b and sections 1.2–1.4, Volume 2): (i) type A3 × × / ×; (ii) type B1 × / × /; (iii) type C × / / ×; (iv) anacrusis added at the head of the verse types that otherwise begin with a lift (type A1 with anacrusis: × / × / ×; type D with anacrusis: × / / × ×; type E with anacrusis; × / × × / (anacrustic verses are all treated as a single set). The length of the verse-initial weak position may be measured and compared in terms of the varying proportions of its minimal (monosyllabic) to nonminimal (polysyllabic) realizations in the three distinct portions of the Old English Genesis, that is, Genesis A1, Genesis B, and Genesis A2. One obtains the following distributions (monosyllabic occurrences/polysyllabic occurrences): for type A3, 3/21 (Genesis A1), 2/47 (Genesis B), 37/142 (Genesis A2); for type B1, 28/104 (Genesis A1), 20/290 (Genesis B), 151/505 (Genesis A2); for type C, 37/83 (Genesis A1), 45/157 (Genesis B), 209/391 (Genesis A2); for anacrusis, 4/4 (Genesis A1), 46/48 (Genesis B), 36/13 (Genesis A2). Performing a Fisher test on the contingency tables devised for each pair of the three portions of the Old English Genesis according to each verse configuration leads to results as follows: for type A3, p = 0.3223 between Genesis A1 and Genesis B, p = 0.4247 between Genesis A1 and Genesis A2, and p = 0.0048 between Genesis A2 and Genesis B; for type B1, p < 0.001 between Genesis A1 and Genesis B, p = 0.7330 between Genesis A1 and Genesis A2, and p < 0.001 between Genesis A2 and Genesis B; for type C, p = 0.1121 between Genesis A1 and Genesis B, p = 0.4606 between Genesis A1 and Genesis A2, and p < 0.001 between Genesis A2 and Genesis B; for anacrusis, p = 1 between Genesis A1 and Genesis B, p = 0.2215 between Genesis A1 and Genesis A2, and p = 0.0072 between Genesis A2 and Genesis B. To generalize, except for type B1, the size of the initial weak position does not differ significantly between Genesis A1 and B, whereas it is consistently distinguished between Genesis A2 and B. In contrast, there is no statistically significant difference detected throughout between Genesis A1 and A2. Reformulated in more revealing terms, the

1.1 The Old English Genesis, the Old Saxon Genesis, and MS Junius 11 

 11

verse-initial weak position is more likely to be realized polysyllabically in Genesis B than either portion of Genesis A. While not directly distinguishable, the same metrical position may be characterized as relatively longer in Genesis A1 than Genesis A2 by virtue of their varying deviation from Genesis B with its marked predilection for polysyllabicity. Thus, insofar as no statistical difference is demonstrably in evidence between Genesis A1 and A2, the situation concerning the relative size of the verse-initial weak position is admittedly not as conclusive as the likelihood of double alliteration adduced earlier in substantiating the integration of Genesis A1 and B in distinction from Genesis A2. Nonetheless, it remains hardly disputable that Genesis A1 is closer to Genesis B than is Genesis A2 in its treatment of the verse-initial weak position, a further manifestation of the convergence between Genesis A1 and Genesis B in divergence from Genesis A2. It is not a new finding at all that Genesis A differs between its earlier and later parts, although its division at the specific point as proposed above, and the pictorial and metrical justifications for the claim as adduced above, are original to the present work. Specifically, the contention that, rather than homogeneous throughout, Genesis A is a composite work consisting of distinct pieces was propounded by Sievers (1929; Menner 1951; Doane 2013: 50, 57). Sievers (1929) divided Genesis A into Genesis I (1–234, corresponding to Genesis Aa in the above exposition) and Genesis III (852–2936, corresponding to Genesis Ab above). The bifurcation is based, plausibly but not conclusively, on an array of linguistic evidence, showing more frequent occurrences of Early West Saxon forms in Genesis I (Doane 2013: 50), and, allegedly and questionably, on Sievers’s idiosyncratic theory and methods of sound analysis (Schallanalyse; Sievers 1924; Ganz 1978: 65–85; Pope 1998: 184–189), which purported to differentiate personal voices and thereby identify distinct individuals’ involvement in poetry composition. While Sievers viewed the two constituents—Genesis  I and Genesis III—as independent in origin, it seems more feasible to ascribe the differences observed to varying editorial interventions exercised on an originally single piece of poetry, Genesis A, as does Doane (2013: 50), for example. Although the division proposed here and Sievers’s are predicated on entirely different grounds, it is interesting to note that they largely overlap, with the exception of the portion after Genesis B, ranging from verse line 852 through verse line 966. It is definitely a worthwhile research topic to evaluate the two partially complementary views—Sievers’s and mine—and work out empirical implications of their disagreements as well as agreements.11 In any event, it should be emphasized that the relative autonomy encompassing Genesis A1 and Genesis B as proposed here—in distinction from Sievers’s view—is empirically motivated and theorized as a reflection of the underlying illustrated Old Saxon Genesis episodes, which in turn would have determined the exact scope of intensive editorial exercise resulting in the Old English/Old Saxon Genesis (1–966).

11 Such empirical assessments will be made in part in section 12.8 below.

12 

 1 Introduction

Table 1 provides an overview of all illustrations in Junius 11 according to their locations in terms of section, page/line, and adjacent verses. With respect to the first twenty-two pictures (corresponding to verses 1a–966b), their constituency in terms of component pictures, where applicable, is noted for ease of reference in the following discussion. On the basis of Doane (2013: 10–21), Table 2 offers section-based accounts of the relevant portion of the Old English Genesis (1a–966b)—the Old English/Old Saxon Genesis, as designated above—regarding their delineation in terms of page/line and verse, pictures contained, and codicological status.

Table 1. Location of pictures relative to section, page/line, and verse  

Subpicture (relevant to A1/B)

Section*

Page/Line ­preceding the picture

Verse preceding the picture**

Verse ­following the picture**

Genesis A1/B/A2

Pii

 

ii/0, full page

— 

1a, p. 1, §(1)

A1

P2 P3

2/12 3/0, full page

48b — 

—  49a, p. 4

A1 A1

P6 P7 P9 P10

  P3.1, P3.2a, P3.2b, P3.3   P7.1, P7.2    

Frontispiece (1) (1) (2) (2) (4) (4)

6/15 7/0, full page 9/11 10/14 (over­ lapping picture)

    P16.1, P16.2 P17.1, P17.2 P20.1, P20.2     P31.1, P31.2 P34.1, P34.2 P36.1, P36.2 P39.1, P39.2 P41.1, P41.2

(4) (5) (6) (6) 7  (11) (12) (12) (13) (13) (14) (14)

11/3 13/11 16/15 17/15 20/0, full page 24/13 28/12 31/0, full page 34/1 36/2 39/11 41/0, full page

P44 P45 P46 P47 P49 P51 P53 P54

  P45.1, P45.2 —  —  —  —  —  — 

16 16 16 16 17 17 18 18

44/0, full page 45/10 46/14 47/16 49/7 51/0, full page 53/0, full page 54/0, full page

938b, p. 43 951b 971b 993b 1035b 1062b, p. 50 1100b, p. 52 — 

—  135a, p. 8, §(3) 186a, p. 10 206a, p. 11 (preceding pages lost) 210a, p. 12 246a, p. 14, §(6) 309b, p. 17 325a, p. 18, § 7 389a, p. 21, §(8) 491a, p. 25 599a, p. 29 663b, p. 32 731b, p. 35 765a, p. 37 842a, p. 40 872a, p. 42, §(15) 939a, p. 45 952a, p. 46 972a, p. 47 994a, p. 48 1036a, p. 50 1063a, p. 52 —  1101a, p. 55

A1 A1 A1 A1

P11 P13 P16 P17 P20 P24 P28 P31 P34 P36 P39 P41

134b —  185b 205b (following pages lost) 209b 245b 309a 324b 388b, p. 19 490b 598b 663a, p. 30 731a 764b 841b 871b, p. 40

A1 B  B  B  B  B  B  B  B  B  B  A1 A1 A1 A1 A2 A2 A2 A2 A2

1.1 The Old English Genesis, the Old Saxon Genesis, and MS Junius 11 

 13

Table 1. (continued)  

Subpicture (relevant to A1/B)

Section*

Page/Line ­preceding the picture

Verse preceding the picture**

Verse ­following the picture**

Genesis A1/B/A2

P56 P57 P58 P59 P60 P61 P62 P63 P65 P66

—  —  —  —  —  —  —  —  —  — 

18 18 18 19 19 19 19 19 20 20

56/10 57/11 58/6 59/2 60/1 61/0, full page 62/7 63/0 65/11 66/9

1142b 1158b 1166b 1178b 1196b 1214b, p. 60 1224b 1236b, p. 62 1313b 1326b

A2 A2 A2 A2 A2 A2 A2 A2 A2 A2

P68 P73 P74 P76 P77 P78 P81 P82

—  —  —  —  —  —  —  — 

21 23 23 24 24 24 25 25

68/1 73/12 74/0 76/17 77/0, full page*** 78/10 81/4 82/14

1362b 1496b —  1584a —  1601c 1696b 1718b

P84 P87 P88

—  —  — 

(26) (27a) (27a)

84/6 87/0, full page 88/0, full page

1766b part of 1830a — 

1143a, p. 57 1159a, p. 58 1167a, § 19 1179a 1197a 1215a, p. 62 1225a 1237a 1314a, p. 66 1327a, p. 67, § 21 1363a, p. 69 —  1497a —  1588b, p. 78 1602a 1697a, p. 82 1719a, p. 83, §(26) 1767a, p. 85 —  part of 1830a, p. 89

A2 A2 A2 A2 A2 A2 A2 A2 A2 A2 A2

* The section numbers in parentheses mean that section breaks are unnumbered in MS. ** When the verse appears at a page/section different from the picture it immediately precedes or follows, its page/section number is given as well. *** Verses 1584b–1588a overlap P77 in the middle.

14 

 1 Introduction

Table 2. Constituency of sections 1 through 16 in terms of pages/lines, verses, pictures, and their status of preservation Section*

Status

Page/Line

Verse line

Picture

(1) Complete 1/1–4/24 1–81 P2, P3 (2) Complete 5/1–6/15 82–134 P6, P7 (3) Lost end 8/1–25 135–168 … —  6 pages lost in between, included in sections (3) or (4) (4) Lost start 9/1–10/14 … 169–205 … P9, P10 2 pages lost in between, included in section (4) (4) Lost start 11/1–12/18 … 206–234 P11 4 pages lost in between, included in section (5); beginning of Genesis B (5) Lost start 13/1–11 … B235–245 P13 (6) Complete 14/1–17/15 B246–324 P16, P17 7  Complete 18/1–19/25 B325–388 P20 (8) Lost end 21/1–22/26 B389–441 … —  4 pages lost in between, corresponding to sections (9) or (10) (11) Complete 23/1–26/21 B442–546 P24 (12) Complete 26/22–32/17 B547–683 P28, P31 (13) Complete 32/18–38/17 B684–820 P34, P36 (B791–817a = OSG 1–26a) (14) Complete 38/19–40/8 B821–851; P39, P41 40/8–40/22 A852–871 (15) Complete 42/1–43/9 872–917 —  16 Complete 43/10–48/6 918–1001 P44, P45, P46, P47 * As in Table 1, the section numbers in parentheses denote that the corresponding section breaks are unnumbered in MS.

1.2 The Cotton Genesis family and the Touronian Bibles As stated at the outset of this chapter, the major contention of this book in the diachronic dimension is that the initial twenty-two pictures in the Old English Genesis are largely derived from the Cotton Genesis family in general and its Touronian branch in particular. A few introductory notes on these manuscript affiliations are therefore appropriate at this point. The Cotton Genesis is a Greek illuminated manuscript of the Book of Genesis, produced in the fifth century, now held at the British Library (Cotton MS Otho B VI; Weitzmann/Kessler 1986). The Cotton Genesis demonstrably served as sources in the Latin West not only for a host of subsequent illustrated biblical manuscripts containing or relating to the Book of Genesis, but also other works of art such as the mosaics and ivory plaques dealing with the story of Genesis. These later works, regardless of their medium, that were crucially inspired by the Cotton Genesis are broadly referred to as members of the Cotton Genesis family. On the basis of the varying extent to which these derivative works are identified as reflecting the Cotton Genesis archetype, Weitzmann and Kessler (1986: 17–29) categorized the family members into three groups—primary,

1.2 The Cotton Genesis family and the Touronian Bibles 

 15

secondary, and tertiary—in order of decreasing fidelity with which the archetypal features are presumably reflected in them, and hence of decreasing evidential value in reconstructing the archetype (for details, see section 12.4).12 Among the manifold family members, the following works figure prominently for comparative examinations of the Junius 11 pictures in this book: 1. The San Marco mosaics, Venice: 13th century; Demus 1984; Büchsel/Kessler/Müller 2014; Weitzmann and Kessler’s (1986) primary group. 2. The Touronian Bibles, Tours; 9th century; Köhler 1930, 1933; Kessler 1977; Weitzmann and Kessler’s (1986) secondary group (for details, see below). 3. The Millstatt Genesis, a Middle High German verse paraphrase of the Genesis, with illustrations; Kärntner Landesarchiv, MS 6/19; 12th century; Millstätter Genesis 1967; Hamano 2016; Schäfer 2019. Weitzmann and Kessler’s (1986) secondary group. 4. The Salerno ivory plaques; Salerno, Museo del Duomo, late 11th century; Bergman 1980; Dell’Acqua et al. 2016; Weitzmann and Kessler’s (1986) secondary group. 5. The Hortus Deliciarum; Strasbourg, Bibliothèque de la Ville (destroyed); 12th century; Green 1955; Green et al. 1979; Weitzmann and Kessler’s (1986) tertiary group. 6. The San Paolo fuori le Mura frescoes (destroyed);13 5th century; watercolor reproductions in the 17th century, Vatican, Biblioteca Apostolica, Cod. Barb. lat. 4406; Waetzoldt 1964. As mentioned above, the Touronian Bibles belong to the secondary group of the Cotton Genesis family. Based on the extant materials, it is generally believed that at least seven illustrated (full) Bibles must have been produced at Tours in the mid-ninth century (Kessler 1977: 8), as listed below (Kessler 1977: 5–8). The first three items are Bibles standing on their own and the remaining four are each inspired by a Touronian Bible, now lost, distinct from any of the three existing volumes. The first four Bibles in the list below, including the Touronian base of San Paolo, are of central importance in the following comparative investigation. The frontispieces to Genesis—the Adam and Eve cycle—in the Touronian Bibles and the corresponding scenes in the Touronian-Bible-based works are largely affiliated with the Cotton Genesis tradition. More specifically, the frontispieces to Genesis in the Touronian Bibles would have been based on a Cotton Genesis copy or something close to it. Accordingly, they are categorized as belonging to the same

12 According to Weitzmann and Kessler’s classification, the Old English Genesis is affiliated with the tertiary group. It should be remembered that they naturally addressed the whole set of forty-eight pictures. The exclusive concern with the twenty-two subset pictures would necessitate radical revision of this categorization, as will be substantiated in this book. 13 Unlike the preceding five sources, this work is dated earlier than the Cotton Genesis (Weitzmann/ Kessler 1986: 17).

16 

 1 Introduction

family, along with the works mentioned above in connection with the Cotton Genesis family: 1. Bamberg (or Alcuin) Bible: Bamberg, Staatsbibliothek, Msc. Bibl. 1; Köhler 1933: 102–105; Fischer 1957; Kessler 1977; two miniatures. 2. Grandval (or Moutier-Grandval) Bible: London, British Library, Cod. Add. Ms. 10546; Köhler 1933: 13–27; Bibel von Moutier-Grandval 1971; Kessler 1977; four miniatures. 3. Vivian Bible (or the First Bible of Charles the Bald): Paris, Bibliothèque Nationale, Cod. lat. 1; Köhler 1933: 27–65; eight miniatures. 4. Bible of San Paolo fuori le Mura: Rome, San Paolo fuori le Mura; Kessler 1977; Biblia sacra 1993; Koehler/Mütherich 1999: 109–174; Cardinali 2009; seven of its twenty-four frontispieces are based on a Touronian Bible: 1r, 3v, 8v, 31v, 170v, 259v, 310v. 5. Bronze door of St. Michael’s Church, Hildesheim; 11th century; Nordenfalk 1971; Kessler 1977; Kaspersen 1988; based on an early lost Touronian Bible. 6. A fresco at the Abbey church of Münstereifel; 12th century; Kessler 1977: 6, fig. 110; based on a Touronian Bible that was preserved at the Abbey, Prüm, during the Middle Ages. 7. Fragments of a Touronian Bible; Nordenfalk 1936: 281–295; Kessler 1977: 6; preserved at St. Maximin, Trier, during the Middle Ages. It is assumed, following Kessler (1977), that the four pandects—including the Touronian antecedent of the San Paolo Bible—were based on related (but not necessarily the same) models; each of these Bibles selected and adapted the diverse sources available on its own and in its distinct ways. As far as the Genesis illustrations are concerned, there would have been a common set of model pictures of the Cotton Genesis origins—a Genesis manuscript of the Cotton family—which would have contained at least nineteen scenes of the Adam and Eve cycle (Kessler 1977: 140). Since the Junius artist is demonstrably indebted to all four Touronian Bibles without appreciable distinction in degrees of influence and preference,14 I will treat as given a whole range of variation observed among the four cognate works, without exploring synchronically the system of their interrelationship, and diachronically their origins and development in the Touronian tradition in general, and their relative and absolute chronology in particular, as has been investigated vigorously in earlier scholarship such as Köhler (1933), Nordenfalk (1936), Gaehde (1971), Kessler (1977), Klein (1984), Cook (2007), to name but a few.

14 On the designation “Junius artist,” see section 1.4.

1.3 Raw (1953; 1976) on Carolingian sources of Junius 11 pictures 

 17

1.3 Raw (1953; 1976) on Carolingian sources of Junius 11 pictures The claim proposed here that a subset of Junius 11 pictures is derived from the Touronian Bibles is in fact far from new. It has been in currency well over half a century since Raw (1953) originally put forward the idea. Having classified all Junius pictures into three groups on the basis of their distinct sources, Raw (1976: 127)—Raw (1976) is a condensed, published version of Raw (1953)—postulated the Carolingian heritage for the main group consisting of the following four series, the Fall of the Angels, Adam and Eve, Cain and Abel, and Noah and Abraham, that is, Pii, P2–P3, P9–P51, P60–P61, and P65–P88. More specifically, Raw (1976: 146–148) inferred that these Carolingian, mainly Touronian, drawings originated from an illustrated Old Saxon Genesis, produced in the mid-ninth century at Tours. Brought to England not long after production, this illuminated manuscript would have been used as a model for the Junius drawings by the Anglo-Saxon artists around 1000. While apparently quite similar in overall terms, Raw’s thesis and mine differ radically in a wide spectrum of specific analyses, assertions, and implications, as well as problematics addressed and major contentions proposed. First, as should be immediately clear from the above paragraph, Raw’s claim is more inclusive than the present one, which is exclusively concerned with the smaller set ranging from Pii to P46, a restriction that has been justified in section 1.1 above, in part by the pictorial and metrical autonomy of the restrictive set in question, and will be fully substantiated throughout this book. No lesser difference is attached to the term “Tours/Touronian” by which Raw and I mean two related but dissimilar notions. Specifically, with “Tours” Raw has in mind the provenance of the original illustrated Old Saxon Genesis, and when referring to sources of influence, she invokes a wider notion “Carolingian” instead, as in Raw (1976: 146): “From all this a Carolingian ancestor for the third, main group of illustrations in the Junius manuscript can safely be inferred.” By contrast, I use the term “Tours/ Touronian” as an art historical reference to the School of Tours, one of the many schools on Carolingian soil. Thus, my claim is formulated in stricter and narrower terms than Raw’s. Accordingly, in her conceptually less circumscribed framework, the derivation of Pii and P2 in Junius 11 from the Utrecht Psalter (Utrecht, Universiteitsbibliotheek, MS 32) by no means contradicts or weakens her identification of Junius pictures as Carolingian in heritage. By my more stringent conceptualization, however, this generalization is contradictory and unacceptable, as is argued below. Second, Raw is primarily concerned with what the Anglo-Saxon artists would have done with the illustrated Old Saxon Genesis, which she tended to deify as an absolute given, rather than treating it as an object of critical inquiry. By contrast, I am principally engaged in what Raw left as it was, that is, what and how the original Touronian artist would have conceived, evaluated, and executed in his project of illustrating a series of Old Saxon Genesis episodes, and, where relevant, what consequences his artwork would have brought about on the Cotton Genesis tradition and beyond. To rephrase rather schematically, Raw addressed the Junius pictures as they came into being in the making

18 

 1 Introduction

of the Junius manuscript (or its Anglo-Saxon antecedent), whereas I conceptualize them as reflections of the Touronian originals and accordingly treat them as material for reconstructing the initial Old Saxon Genesis illustrations. Third, one of the major pieces of evidence that Raw (1976: 146–147) adduced for the Touronian origins of the Junius pictures is the closer correspondence in terms of content that she claimed to exist between the pictures and the Genesis B text than between them and the Genesis A counterpart. In contrast, my thesis rests on no such text–image correspondence or lack thereof for justification. Concretely, the detailed narrative of the Fall of the Rebel Angels provided in Genesis B is allegedly better reflected in the illustrations, P2 and P3. Accordingly, Raw reasoned, the illustrations at issue originally would have been associated with an Old Saxon text that underlies Genesis B. As Broderick’s (1978: 403–412; 2009: 385–387) detailed critique has shown persuasively,15 Raw’s argument proved to be inconclusive. Given frequently conflicting or ambivalent text–image relations observed, one cannot confidently decide on Genesis B over Genesis A as an optimal fit for the extant images. Fourth, and most substantively, Raw’s thesis depends for its credibility on her specific proposals (Raw 1976: 139–146) for deriving the Junius pictures and/or their constituent detail—particularly those belonging to the Adam and Eve cycle in her view16—from the Genesis frontispieces in the Touronian Bibles, analyses with which I mostly disagree, and as alternatives to which I submit more explanatory accounts in the subsequent chapters. Most conspicuously, the irreconcilable divergence between Raw and the present author is symbolized by our conflicting views on the sources of Pii (God Enthroned, Pl. 1) and P2 (God Enthroned, Pl. 2). Raw (1953: 68–70) derives them from the Utrecht Psalter and the Roda Bible (Paris, Bibliothèque Nationale de France, Ms. lat. 6 [3]), respectively, in either case outside the Touronian Bibles. By contrast, privileging the two pictures as the paired prime members of the system of pictorial organization for the twenty-two pictures of Junius 11 (section 12.1), I identify the Vivian Bible, fol. 423r (Dedication frontispiece, Pl. 32), as underlying both, which are characterized as variant reflections of the archetype embodied in Vivian 423r (section 2.2). Raw’s interpretations rightfully induced Broderick (1978: 470–475) to raise doubts and objections as to their plausibility and adequacy.17 Careful evaluation of Raw’s and my mutually conflicting analyses and interpretations—particularly concerning the identification of specific Touronian models for the Junius pictures—will be explored at length where appropriate in individual sections that follow.

15 See also Lucas’s (1981: 6–12) criticism of Raw’s thesis. 16 “The clearest signs of Carolingian influence on the illustrations of Junius 11 are found in the pictures of Adam and Eve (pp. 9–46)” (Raw 1976: 139). 17 “The fact of the matter is that there is not one single detail in all of the Junius illustrations that would indicate that these illustrations had to have been based on a Carolingian manuscript of the Tours–San Paolo type” (Broderick 1978: 471).

1.4 Notes on notations and referring expressions 

 19

Elaborating on the preceding point relating to possible source material in the Touronian Bibles for comparative investigation, while Raw confined herself to the Genesis frontispiece for whatever unstated reason, I do not impose such a pretheoretical, conceptual restraint in advance. On the contrary, in light of rich potential resources available for image creation, I will remain open-minded to non-Genesis pictures amply contained in the Touronian Bibles, especially the Vivian and San Paolo Bibles. This conceptually wider and methodologically flexible perspective on feasible source material naturally makes possible formulation of alternative accounts and postulation of derivational mechanisms that were categorically inaccessible in Raw’s conceptual framework. Fifth, and closely associated with the previous point, unlike Raw, I will be problematizing squarely the interpictorial relationship in Junius 11, under the working hypotheses that all pictures are, synchronically and statically seen, integrated into the network through mutual interdependence of various kinds, and that, diachronically and dynamically seen, the network provides a structural basis for constructing novel images. Raw did not bring in such a structural conceptual framework, at least in an explicit manner. This formal–structural orientation will open up a new vista for research by making possible posing questions otherwise inconceivable and seeing meaningful relations where none were even imagined to be in existence. Sixth, in proposing the Touronian origins of the Junius drawings, Raw subsumes not only the Touronian Bibles but also the Utrecht Psalter and works affiliated with the Court School of Charles the Bald under too broad a notion of Carolingian influence (e.g., Raw 1976: 146), as noted above. While this general nomenclature may be useful as a means of stressing the continental inheritance of the Junius pictures as opposed to an insular tradition—the kind of problematics Raw defined for her study—such a vague notion does more harm than good for the research agenda pursued in this book. More specifically, the practice of lumping together the Utrecht Psalter with the Touronian Bibles is totally unacceptable, since whatever inspiration the Utrecht Psalter would have given to the Old Saxon Genesis pictures cannot be compared remotely with the foundational contributions made by the Touronian Bibles, as demonstrated throughout this book and consolidated in Chapters 12 and 13.

1.4 Notes on notations and referring expressions The portion of the Old English Genesis text that has been delimited (verse lines 1–966) as corresponding to the pictorial narrative from Pii to P46 is referred to as the Old English/ Old Saxon Genesis for lack of a better name. Genesis A1, introduced in section 1.1 above, covers only the subpart of this whole complex (verse lines 1–234 and 852–966), excluding Genesis B. For ease of reference, the special notation for a Junius picture will be used in the form of P13, for example, which identifies the illustration on page 13 in the Junius 11 manuscript, whereas page 13 (or p.  13) stands for the manuscript page physically conceived per se. To facilitate search with maximum efficiency, each Junius

20 

 1 Introduction

picture is individually referred to by prefixing P before its page number, even when multiple references are made, such as P10 and P11 (rather than PP10 and 11), P34/P36/ P39 (rather than PP34/36/39). Further, where a Junius picture page consists of two constituent pictures by a vertical division of the space between the upper and lower halves with a distinct boundary in between, the distinguishing number 1 or 2 (ordered from top to bottom, not always corresponding to narrative sequence) is supplied for individual identification, as with P31.1 (the upper half of P31). In similar fashion, register-based pictures such as Bamberg 7v are differentiated by the same notation augmented, where necessary, with a further distinguisher regarding the horizontal location within a given register: for example, Bamberg 7v3 (right) stands for the right scene in the third register of the composite pictorial representation in folio 7v of the Bamberg Bible. Notice in this connection that unless referential ambiguity arises, each illustration is referred to by a short transparent name, followed directly without a comma by a folio number (where applicable), as with Vivian 423r for the Vivian Bible, folio 423r, and Bamberg 7v for the Bamberg Bible, folio 7v, given immediately above. Unless otherwise noted, the terms pre-Junius or Junius artist/illustrator/illuminator and the like denote the original artist who would have been responsible for composing maximally the twenty-two pictures (or a core subset of them) underlying those found in the extant Junius 11 manuscript, Pii–P46. According to my conceptual framework, the individual involved was a Carolingian artist at Tours, rather than an Anglo-Saxon illuminator or copyist working in England, as will be synthesized in section 12.8.

2 God Enthroned: Pii/P2, with a brief discussion on P3.2b, P16.1, and P17.1 2.1 A synchronic–structural perspective The frontispiece Pii (Pl. 1) is a full-page representation of God enthroned. However, it displays a number of uncommon features in deviation from typical instantiations of the motif. First is the absence of a mandorla or a circle enclosing the Creator, a common attribute of God in Majesty, as actually depicted in P17 (Pl.  11). Instead of the usual mandorla or circle, the whole image including the attending seraphim is framed by an angled arch, a unique feature scarcely encountered elsewhere, as pointed out by Broderick (1978: 58–59). Apparently similar is the Carmen Paschale (Antwerp, Plantin-Moretus Museum, M 17.4, fol. 1r; Alexander 1978: fig. 285), although the framing arch in it is rounded. Second, with respect to the illustration of the Deity, the left half of the body deserves particular attention, as it suffers a series of deformations. The left hand holds what is often described as a scepter or a scroll (hereinafter referred to as a scroll/scepter), whose upper portion above the grasping hand is banded, while the lower half is left plain. That the identity of the object remains unsettled largely depends on its atypical length and width: it seems too long and thin for a scroll, and too short and thick for a scepter.18 A further anomaly concerning the Creator’s depiction is the segment immediately below this scroll/scepter. This vertically long square is obviously distinct from the scroll/ scepter located immediately above it, as it is larger in width and differs in shape. This object is open to two conflicting interpretations. First, it may be identified as a codex standing on the left knee, as explicitly depicted immediately following the Apotheosis of Boethius (Paris, Bibliothèque Nationale, Ms. Lat. 6401; Temple 1976: 59), fol. 158v (Pl. 73; Temple 1976: pl. 94). In this view, the scroll/scepter is placed on top of the codex, as in Boethius 158v. In contrast to the Boethius representation, however, the hand does not hold the top of the codex and the bottom of the scroll; rather, it grasps the middle of the scroll/scepter, which in turn supports the codex from above, a rather unnatural and unstable configuration. Second, the rectangular portion that lies under the bottom of the scroll/scepter may be regarded as the left leg of the Creator. In other words, the bottom of the scroll/scepter is placed directly on the lap without an intervening item. As remarked by Broderick (1978: 62), the left knee thus identified seems too large in size in comparison with the right. If the scroll/scepter were longer, thereby occupying much of the area at issue, the left knee and leg correspondingly would be positioned lower

18 For various interpretations advanced to date, see Broderick (1978: 56n2). For a derivational trajectory of this enigmatic object, see the following section. https://doi.org/10.1515/9783110788068-002

22 

 2 God Enthroned: Pii/P2, with a brief discussion on P3.2b, P16.1, and P17.1

in harmony with its right counterpart. Thus, both interpretations incur an unnatural configuring of relevant elements. Of further importance, in addition to the unusual posture assumed, either of the above two interpretations appears at variance with the drapery that is depicted extending from the bottom of the scroll/scepter downward to the right. In a more natural composition, the top of the drapery should be relocated upward to the level of the gripping hand, as illustrated in Vivian 329v (Pl. 30), San Paolo 259v (Pl. 36), Dufay Gospels (Paris, Bibliothèque Nationale, Ms. lat. 9385), fol. 179v (Köhler 1930: pl. I.109a; Kessler 1977: fig. 62), Le Mans Gospels (Paris, Bibliothèque Nationale, Ms. lat. 261), fol. 18r (Köhler 1930: pl. I.118b; Kessler 1977: fig. 63), Lothair Gospels (Paris, Bibliothèque Nationale, Ms. lat. 266), fol. 2v (Köhler 1930: pl. I.98b; Kessler 1977: fig. 61), and Boethius 158v (Pl. 73). Alternatively, the lower unbanded portion of the scroll/scepter should be redrawn as part of the drapery, as in Boethius 158v. Situated in the current position, the puff possibly might be viewed as erroneously representing the tapering part of one of the seraph’s wings, in light of the comparable illustration in P2 (Pl. 2), in which the upper portion of the drapery is covered with feathers (for details, see section 2.3 below). Addressing squarely the peculiar drapery in question, an alternative interpretation may regard it as an attribute of the left hand as it is, without postulating relocation or redrawing. In the existing illustration, the left hand itself is not depicted, to be sure, but its presence readily may be inferred on the strength of numerous illustrations such as Vivian 329v (Pl. 30), San Paolo 259v (Pl. 36), Dufay Gospels 179v, Lothair Gospels 2v, and Boethius 158v (Pl. 73). This reinterpretation, far from self-evident as it may appear, correspondingly requires reconsideration of the hand grasping the scroll/scepter: no longer identifiable as belonging to the Creator, the hand must be reinterpreted as another being’s. Since there is only one being in sight on the right side of God, the hand must belong to the seraph.19 In fact, such an attribution appears to be confirmed by the drawing in P2, in which the gripping fingers—under one reading—spring from the lower left wing, which tapers off into the drapery. The seraphim, however, usually do not hold anything but spread their hands open, as exemplified in San Paolo 117r (Pl. 35), Boethius 158v (Pl. 73), and so on. This resultant extraordinary composition will prove to be of some significance in comparative and diachronic terms, as shown in the following section. The enthroned Deity is surrounded by four faces, two being situated immediately above God’s hands (according to one interpretation of the left hand) and the other two just under the feet of the throne. Of particular interest about these faces is their outward orientation, which may be counted as a third peculiarity in representation: they are all

19 This reinterpretation may have been promoted by the model (Vivian 423r, Pl. 32) in which the two attendants to the king’s right hold a rod-like object: the retainer grasps the frame of the back of the throne, and the bodyguard holds a sword. For details, see the following section.

2.1 A synchronic–structural perspective 

 23

turned away from God, looking outside. In other comparable illustrations of God/Christ in Majesty, however, the standing seraphim look inward at the Creator, as in Vivian 130v (Pl.  28), San Paolo 117r (Pl.  35), and Boethius 158v (Pl.  73). An exception to this generalization is the St. Gauzelin Gospels (Nancy, Trésor de la cathédrale, Cathédrale de Notre-Dame de l’Annonciation), fol. 3v (Köhler 1930: pl. 35c; Hubert/Porcher/Volbach 1970: pl.  120), similar to the Bamberg Bible 339v, which lacks seraphim, however: in St. Gauzelin, the two flanking seraphim look outward, as in Pii. Most remarkably, P2, a close variant of the frontispiece under consideration, represents the contrary default configuration of two seraphim looking inward at the Creator. Of vital significance for understanding specific details of Pii including the array of anomalous features pointed out above is P2 just mentioned. These two drawings are strikingly similar in numerous respects with regard to the representation of the Creator (Broderick 1978: 75): they both depict the Deity seated on a backless throne, holding a scroll/scepter in its middle portion in the left hand, raising his right hand with a pointing gesture, and keeping the left leg raised higher than the right. In fact, as testified in the accompanying inscription “hǣlendes hēhseld” (the throne of the Savior), this picture represents God enthroned, as does the frontispiece, but from a radically different perspective and with a marked shift in focus. Correspondingly, there is indisputably an array of notable dissimilarities between the two drawings, as examined at length in the following. The two representations of the enthroned Deity differ from each other along a number of compositional parameters. Their variations, however, are far from random individual differences, rather they derive structurally from the underlying distinction in perspective and interest that motivates the two respective drawings, as will be shown below. Identifying a set of structural oppositions involved between the two pictures will lead to their explanatory accounts in mutually illuminating and reinforcing ways. As far as framing is concerned, a few parametric differences may be recognized. First, the overall framing is symmetrical and balanced on the vertical axis in Pii, whereas it is asymmetrical and unbalanced in P2, in which the double-lined frame is provided fully on two sides (top and right), only marginally on the upper left, and none at all on the bottom. Given that Pii also lacks a boundary at the bottom, one may be concerned with the existing three sides and the extent to which they are materialized. Second, the framing is elaborately designed in the form of an arch and columns in Pii, but it is simple—double outline—in P2. Third, all representations are properly contained within the frame in Pii, while this does not hold in P2, where the tips of the upper wings of the seraphim extend over the top border of the frame. As regards the overall treatment of the framed space available for drawing, two contrastive features are noteworthy. First, the figures in Pii—the Creator, the seraphim, and the cherubim—fill the entire space enclosed by the frame, and feature centrally as an overall symmetrical configuration. The corresponding beings in P2—the Creator, the seraphim, and two other angels—are represented on the upper right periphery with a larger space below left blank, and arranged in an asymmetrical manner: on the right of

24 

 2 God Enthroned: Pii/P2, with a brief discussion on P3.2b, P16.1, and P17.1

the Deity stands a four-winged, limbless seraph alone, while on the opposite side three beings are depicted, including the six-winged seraph with hands and feet. Second, while God, represented largest among the constituent figures, occupies the central place in Pii, the smallest angel stands at the center in P2, with the Creator relegated to the right corner. Concerning particular manners of depicting figures, three sets of contrastive properties may be identified. First, Pii and P2 represent the Creator with differing degrees of detail. As a natural consequence of the greater size of the God image, Pii provides a more detailed depiction than P2, as evidenced by the centrally parted hair, the finer beard and whiskers, and the more elaborate illustration of folds in the robe. Conversely, P2 is differentiated from Pii by significantly less-articulated representations of these segments. Second, the orientation of faces is divergent in the two drawings, as observed above. Specifically, God is depicted facing in diverse directions, in frontal (Pii) on the one hand, and in three-quarter view to the left (P2) on the other. The attending angelic beings look outward in Pii and inward to God in P2. Put another way, while the eyes are all turned outward away from each other in Pii, they are directed to one specific figure in the group—the four angels gaze at the Creator, who in turn looks at the smallest angel. Third, the Creator, the seraphim, and the cherubim in Pii are partially overlapped and contiguous in extension without a categorical spacing in between: in other words, the seraphim and cherubim seem to be fused with God into one unified constellation (cf. Broderick 1978: 58, 70). Such a merging, compact configuration has a further consequence for the depiction of the angelic beings: they are without hands and feet, and the heads—the only distinct body part represented—seem to be immersed in the surrounding wings. In sharp contrast, separated from one another, the figures in P2 are highly discrete in shape, individuated in contour, and diffuse in arrangement. In addition, one of the seraphim is fully depicted with hands and feet. Three further contrasts are detected which concern the inverse arrangement of representational attributes. First, the faces of the seraphim are reversed in Pii and P2 not only in their orientation, as pointed out immediately above, but also in substance. More specifically, the face of the left seraph in Pii (in three-quarter view) corresponds to that of the right in P2 (in three-quarter view), and the right seraph in Pii (in profile—notice that, strangely enough, the left eye with the brow is located in the wing; for more on this peculiarity, see section 2.3 below) matches the left in P2 (in profile—notice that what would correspond to the left eye in Pii is incorporated into the upper left wing, which has several other eyes). Second, the eyes and feathers in the wings—the upper two and the lower (inner) two wings—are largely complementary in distribution between Pii and P2. In regard to the left seraphim, while the roots of the lower wings are provided with eyes in Pii, those of the upper wings are comparably equipped in P2. Much the same applies to the right seraph in Pii, in which the corresponding zones of the upper and the lower wings are decorated with feathers and eyes, respectively. P2, however, follows this complementary pattern only in part, as solely the upper right wing, but not the left, is supplied with eyes,

2.1 A synchronic–structural perspective 

 25

and only the lower left wing is feathered but the right remaining blank. The eyes may be characterized as marked in opposition to feathers, by virtue of their complicated design and limited occurrence. In this view, what contradicts the pattern at issue is the absence of eyes in the upper left wing, rather than the absence of feathers in the lower right wing, which must be regarded as an incidental anomaly of being undecorated at all rather than decorated by default with feathers. As will be shown below, the lack of feathers may be related to the weird feathers extending downward from the hand grasping the scroll/scepter.20 Third, represented by a narrow double line, a clavus—a purple stripe on the tunic— is visible either on the right side of the Lord’s upper body (Pii) or the left (P2). Normally, the right-side stripe is visible, with the left one hidden from sight under a toga. The contrastive arrangement seems to lack an appreciable substantive difference in signification, except that the two images are simply differentiated on this parameter. In any event, this property, and the preceding two as well, carry a contrastive value through their complementary occurrence in the pairs of pictorial fields that are distinguished from each other by a binary feature such as ‘upper versus lower’ and ‘right versus left’. The set of oppositional features that have been identified above to be contrastively characteristic of the drawings in Pii and P2 are summarized in Table 3. At this point, several comments are needed on their substantive and relational bases. First, all these features are binary in their value assignment, and logically independent of one another: the value of a given property does not imply a particular value for another property. For example, a symmetrical frame is not necessarily elaborate or simple in design, nor is it fixed in terms of violability of its boundary. Similarly, a space completely enclosed by a symmetrical frame need not be fully used for drawing. Second, and derived from the first point, the specific alignment of parametrical values that obtains in Table 3 is a matter of empirical significance rather than a consequence of logical necessity. Accordingly, the clustering of individual values, such as is represented under columns Pii and P2 in Table 3, requires explanation in empirical terms, rather than regarding the collection as a pure contingency. More specifically, one is challenged to offer credible accounts of why the Pii drawing, rather than the P2 counterpart, manifests the combination of these values, rather than any other logically possible configuration of feature specifications. Third, the feature of framing, which varies between complete and incomplete, may have two differing manifestations: complete enclosure on the one hand and complete containment on the other. Since the two drawings under consideration are framed on 20 The lack of perfect complementariy in the right seraph of P2 may perhaps be connected with the imperfect representation of its wings (four, rather than six), which may in turn be held responsible for the blank area in question. Seen from a different perspective, another possible correlation emerges: while the outer wings of the right seraph are completely blank (zero representation), the inner one is partially blank. These and other observations will be systematically explored to account for such anomalies in section 2.3 below.

26 

 2 God Enthroned: Pii/P2, with a brief discussion on P3.2b, P16.1, and P17.1

three sides excluding the bottom, complete enclosure in the present context should be understood as fully bounded on the remaining three sides. Fourth, unlike the first ten features listed in Table 3, the last three cannot be assigned fixed value in terms of general categorization. At stake here is a purely relational difference: given a pair of contrastive attributes A and B (or simply presence or absence of A, i.e., A vs. –A), and two distinct fields X and Y available for embodying these attributes, A and B (or A and –A) are realized in X and Y, respectively, in Pii, whereas the converse alignment—A and B appearing in Y and X, respectively—obtains in P2. The particular alignment patterns selected (A/X and B/Y vs. A/Y and B/X), however, defy categorization predicated on a conceptual framework and accordingly lack their own inherent meanings. Table 3. Oppositional features of Pii and P2 Feature

Value for Pii

Value for P2

Frame shape Frame design Framing Drawing space Figure depicted at the center

Symmetrical Elaborate (architectural) Complete Fully occupied Largest being (the ruler)

Depiction of God God’s face orientation

Asymmetrical Simple (geometric) Incomplete Partially occupied Smallest being (haloed angel) Simple Three-quarter; facing a particular figure Multiple; two seraphim + 2 Looking inward (toward the ruler)

Detailed Frontal; not facing a particular figure Minimal; two seraphim Looking outward (away from the ruler) Continuous; unified; symmetrical Discrete; individuated; asymmetrical

Number of attendants Orientation of attendants Contour of figures Faces of seraphim (purely relational) Eyes vs. feathers in seraphim’s wings (purely relational) Location of a clavus

A/X; B/Y

A/Y; B/X

A/X; B/Y; consistently patterned

A/Y; B/X; inconsistently patterned

On the right side

On the left side

The contrastive grouping of variables that distinguishes Pii and P2 is an empirical issue that requires explanation, since these specific values are determined independently, as discussed above and their actual alignment was implemented accordingly by exclusion of a wide range of logical possibilities otherwise available. To address such a problem squarely, rather than accepting the existing set simply as given, is tantamount to exploring deeper into the underlying motivations for the construction of the contrastive pair of pictorial representations and thereby reconstituting plausible messages that they were intended to convey. Since the two drawings obviously differ in the extent to which they

2.1 A synchronic–structural perspective 

 27

are susceptible to interpretation at first glance, it is less risky strategically to start with the one seemingly easier to make sense of, namely, Pii. In light of this more tractable image, then, the more recalcitrant one, P2, eventually may become subject to elucidation by analyzing the network of oppositions and complementarities between the two pictures. Identifying Pii as a representation of God enthroned seems to be credible beyond reasonable doubt, despite the array of irregularities exhibited in the drawing, as remarked in the beginning of this section. Such a categorization in turn may lead unequivocally to determining the three features that directly bear on the ways the Creator is represented, as follows: God, the glorious king of hosts, is depicted carefully as a central image, dominating the drawing space available. Furthermore, the seraphim are closely associated with the heavenly king as his attendants, as portrayed in Isaiah 6:1–2 (Karkov 2001: 46),21 in which the prophet offers a visionary account of the Lord seated on the throne at the royal court in the Temple: “. . . I saw the Lord sitting on a throne, high and lofty; and the hem of his robe filled the temple. Seraphs were in attendance above him; each had six wings . . .” (New Revised Standard Version, Coogan 2010).22 No less importantly, the vision of God enthroned on the cherubim is commonly referenced, as described in, for example, 1 Samuel 4:4: “. . . the Lord of hosts, who is enthroned on the cherubim” (Coogan 2010; similarly in Ps. 80:1 and Ps. 99:1), and illustrated—albeit in varying forms—in the Utrecht Psalter (Utrecht, Universiteitsbibliotheek, MS Bibl. Rhenotraiectinae  I Nr 32; DeWald 1932; Dufrenne 1978; Utrecht Psalter 1984; Koehler/Mütherich 1994: 36–38, 85–135; Utrecht Psalter 2015), fol. 59v (Pl.  79; Broderick 1978: 66) and the Bamberg Daniel Commentary (Bamberg, Staatsbibliothek, Msc. Bibl. 22), fol. 5r. Of related interest is the description of the construction of Solomon’s Temple in 1 Kings 6, particularly the Holy of Holies, the inner sanctuary: “In the inner sanctuary he made two cherubim of olivewood, each ten cubits high. [. . .] He put the cherubim in the innermost part of the house; the wings of the cherubim were spread out so that a wing of one was touching one wall, and a wing of the other cherub was touching the other wall; their other wings toward the center of the house were touching wing to wing” (1 Kings 6:23–27; Coogan 2010). Ezekiel 1:4–28 provides a more vivid account—partly complementary to and partly conflicting with 1 Kings—of the cherubim and the enthroned Lord above them. Of central concern is the following vision: “And above the dome over their heads there was something like a throne, in appearance like sapphire; and seated above the likeness of a throne was something that seemed like a human form” (Ezek. 1:26; Coogan 2010; see also Ezek. 10:1). In light

21 Reinhard (2010) claims that a proper understanding of Pii requires an integral reading of Isaiah 6:1–2 in conjunction with the following two verses (Isa. 6:3–4). This widened literary context, he contends, provides an allegedly revealing interpretation of Pii, including its architectural frame and wavy ground. In section 2.2 below, I present more specific and explanatory accounts of these features in art historical terms. 22 For a survey of varying representations of the motif of God seated or standing between two seraphim, see Raw (1997: 133–136).

28 

 2 God Enthroned: Pii/P2, with a brief discussion on P3.2b, P16.1, and P17.1

of these descriptions in the Old Testament, the two-winged faces under the throne in Pii may be reasonably identified as cherubim in agreement with Gollancz (1927: xxxix; cf. Broderick 1978: 65). Of interest in this connection is the similarity in shape—isosceles trapezoid—between the winged being under the throne and the foot of the throne in P2. Given that the two illustrations in Pii and P2 are characterized as variations of the same representation (for details, see below), this geometrical correspondence confirms Gollancz’ interpretation. As with the beard and whiskers of God, the bottom part of the feet of the throne receives detailed representation in Pii, while the corresponding elements are left unarticulated in P2.23 Inasmuch as the seraphim and cherubim are envisioned integrated with the temple—heavenly or earthly—as their grounding location, it is only natural to find some architectural elements of the temple reflected in a pictorial representation of God enthroned, particularly where these attending beings figure prominently, as in Pii. Seen in this light, the more elaborate framing of Pii over P2—the depiction of an arch and columns, and particularly the segmentation of the latter into capitals, shafts, and bases, despite their crude appearances—makes better sense than the contrary alignment, especially when it is noted that P2 does not contain cherubim in its drawing. Despite such a conceptual linkage of the temple with the enthroned God and the seraphim/cherubim in their usual presentations, however, the building does not serve as an indispensable component of a representation of God in Majesty. Therefore, a more specific explanation is needed for the actual occurrence of—rather than merely a greater likelihood of—architectural elements in Pii as opposed to P2, beyond just invoking the metonymic and syntagmatic adjacency. Similarly, the overall integration of God and his attendants into a single unit in Pii cannot be simply ascribed to the compositional centrality of the Lord in the motif, given numerous illustrations of the enthroned God in which the seraphim/cherubim are depicted marginally and separately in sharp distinction from the Deity at the center, as in San Paolo 117r (Pl. 35).24 In addition to the above features, which are only partially explainable, the symmetrical and complete framing and the outward orientation of flanking attendants are hardly subject to a principled account, without taking into consideration the pictorial opposite of Pii, namely, P2. In other words, specifications for these features are deter23 Also interesting is a similar composition of two heads under the feet, which occurs in Vivian 415v, the Apocalypse illustration, in which an angel is represented standing upon two heads, the personifications of the sea and land, partly corresponding to Revelation 10.2: “. . . Setting his [= the angel’s] right foot on the sea and his left foot on the land” (Coogan 2010; cf. Kessler 1977: 79–80). Given the paramount role that Vivian 423r plays in the composition of Pii and P2 in Junius 11 as will be demonstrated in section 2.2 below, the occurrence of a compositionally comparable scene on the immediately preceding illustration page (i.e., Vivian 415v) would have had some bearing on the representation of cherubim in Pii. Compare also the Bamberg Daniel Commentary, fol. 5r, in which two angelic beings support the globe throne on which God is seated. Compare Raw (1997: 135). See Isaiah 66.1: “Thus says the Lord: Heaven is my throne and the earth is my footstool; . . .” (Coogan 2010; cf. Raw 1997: 137). 24 On the symmetrical shape of the consolidated God and angels, see Ohlgren (1972b: 265).

2.1 A synchronic–structural perspective 

 29

mined not on an inherent basis of the underlying motif of Pii, but derivatively through structural opposition to the design organization of P2 that is primarily motivated on inherent grounds.25 As suggested above, P2 is a variant of God enthroned, with a different focus of concern. Unlike Pii, the Creator seated in glory is not imaged as an absolute self-existence. Rather, it is presented in relative terms with a crucial reference to the small angel, who features centrally in the scene. Given the allotted space, this haloed angel may be counted as a major figure in P2, along with the Lord. The question arises: Who is this privileged angel? As is generally held (Gollancz 1927: xxxix; Raw 1976: 146–147; Broderick 1978: 75, 87; Ohlgren 1992: 88), this most plausibly may be identified as Lucifer, the brightest of all angels (cf. Genesis B 252b–256a). The identification is justified on compositional grounds, if nothing else: since this is the only angel in Junius 11 to be provided with a halo (Gollancz 1927: xxxix; Broderick 1978: 74), it must have a correspondingly privileged status among the hosts of angels; hence, it may be inferred that the figure must be Lucifer, God’s most beloved, despite its childlike appearance in this picture, for which a cogent explanation will be offered below. The identification of the smallest angel in P2 as Lucifer is confirmed by the representation in P3 (Pl. 3; section 3.1). In the first two registers (P3.1 and P.3.2a) stands a crowned angel in the center, who is depicted largest in each panel and approached by a procession of other angels paying him homage. No longer represented as a child, the central angel in the top register is doubtless meant to be none other than Lucifer, the leader of the rebel angels: pointing to the throne he claims to be seated on as his exclusive privilege, he poses himself as a new ruler, daring to supersede the Lord. The thematic continuity of the P2 illustration to the P3 pictures26—structurally parallel to the integrity of P6 and P7 (section 4.1)—that the foregoing discussion assumes is supported not only by the common centrality of the angelic beings, but also by the right borderline of P2 that extends farther below the ground line, although the large space to the left remains blank. This excessive length may have been designed as an iconic guide for viewers to thematically unite P2 with the following scenes in P3. Furthermore, the lack of a bottom line (in contrast to the presence of one in P10, for example) might be construed as a supplementary device for signaling the thematic open-endedness of P2 in general and its narrative continuation to P3 in particular. In addition to indicating the narrative unity of P2 and P3, the extended line’s intended function also would have been to delineate the P2 drawing from the facing bottom register in P3 (P3.3) on the right, thereby precluding the blank area and the facing register from being brought together 25 Needless to say, one may appeal for a partial explanation to the contingent fact that God enthroned is usually symmetrically represented on the vertical axis. 26 Consider Ohlgren’s remarks: “This illustration and the upper two registers of the drawing on page 3 [. . .] constitute the visualization of a thematically-important textual episode—the pride, rebellion, and fall of Lucifer” (Ohlgren 1975: 39). See also Ohlgren (1972b: 265–269). For a critique of Ohlgren’s interpretation of P2 formulated in more specific terms, see below.

30 

 2 God Enthroned: Pii/P2, with a brief discussion on P3.2b, P16.1, and P17.1

in interpretation. The use of such a preventive device makes further testimony that P2 and P3 constitute a unified picture series.27 It should be noted in this connection that the opposite left frame line in P2 stops well above the ground line, yet it is long enough to keep the inscription (as mentioned above) from intruding the drawing area. Such a characterization then prompts the following question: What is the content of the P2 illustration? Previous studies generally regard the scene as one of confrontation between God and Lucifer: the Deity is “expostulating” with (Gollancz 1927: xxxix) or “reprimanding” (Raw 1953: 28) Lucifer; “God is meant to be understood as angry” (Broderick 1978: 87). More specific is Ohlgren’s (1972b: 265) account, deserving careful evaluation: The drawing on page 2 […] depicts an interview scene between Lucifer, who presumably has just announced his plans to establish a rival throne in heaven, and the Deity, who appears to have an astonished look on His face.

Ohlgren’s interpretation is in turn built on his textual analysis—inspired by Dorfman’s (1969) narrative analysis of medieval epic—of the fall-of-the-angels episode in the Old English Genesis. According to Ohlgren (1972b: 264), as quoted below, the episode is composed of five units in terms of linear progression of its smaller narrative constituents (incidents): (1) God and the blessed angels live harmoniously in heaven (ll. 1–18a); until (2) Lucifer, through his pride, establishes a rival throne in heaven (ll. 22a–34). God, upon hearing of the rebellion, (3) becomes angry and wrathful (ll. 34b–36a), (4) creates hell (ll. 36b–46b), and (5) drives the rebel host into hell (ll. 47a–76b).

Although Ohlgren (1972b) does not treat 18b–21b (quoted below) in his narrative analysis, these lines should no doubt be dealt with as part of unit 1. Ohlgren (1972b: 269–271) contends further that the episode at issue is repeated twice, at 246a–441 and 733b–763a. synna ne cuþon, // firena fremman / ac hīe on friðe lifdon // ēce mid heora aldor. / elles ne ongunnon // rǣran on roderum / nymþe riht and sōþ (They did not know of sins, the doing of evil deeds, but rather they lived in peace, eternally with their leader. They strove to exalt nothing else but right and truth, …) (Anlezark’s [2011: 3] translation)

Drawing on the above textual analysis, Ohlgren goes on to claim that P2 and P3 constitute the visualization of the second narrative unit, Lucifer’s pride. While P3 readily succumbs to such a characterization, there seems to be nothing specific in P2 that dis-

27 On the significance of the large blank space in P2, see below. Briefly, the blank space arises from removal of other figures from the scene who were not considered to be worthy of depiction for the present purpose, a symbolic representation of the hierarchical relation between the Deity and Lucifer. More importantly, the empty space is designed to signify that hell is not in existence yet, in contrast to the facing register that depicts the subterrestrial realm.

2.1 A synchronic–structural perspective 

 31

plays Lucifer’s arrogance and disloyalty, the defining property of the narrative unit in question. However, this does not disprove Ohlgren’s interpretation, as the picture in itself, open to a variety of interpretations, hardly would contradict an attempt to read into it Lucifer’s denouncement of obedience to the Lord and his resulting astonishment, as Ohlgren (1972b) would have us believe. As far as the texts of Genesis A and Genesis B are concerned, however, one cannot find a place that seems to even remotely suggest a denouncement scene comparable to what Ohlgren takes for presumable; nor need such an event be postulated as indispensable for a coherent understanding of the whole episode of the rebellion and fall of angels. It would not necessarily have been in a face-to-face meeting with boastful angels that God would have learned about their insolence and conspiracy. On the contrary, Genesis B seems to imply that Lucifer tried to keep his treachery hidden, as described in 261b–266a. Although Lucifer spoke many words of insolence (271b–272a, 292a–295a), they would not likely have been made in the Lord’s presence. This is implied in the first two registers in P3—partly embodying what Lucifer boasted about—in which Lucifer exalted himself among his followers, but the Deity is nowhere illustrated. No more compelling to posit would be the scene in which God, upon hearing about the angel’s disobedience, rebuked him in person for his insubordination, contrary to Gollancz’s (1927) and Broderick’s (1978) conjecture. In summary, Ohlgren’s interpretation of P2 as an interview scene between God and Lucifer proves to be insubstantial and far-fetched in light of available evidence, although it remains a possibility, and cannot be rejected outright. Accordingly, one must explore an alternative account that is more congruent with pictorial and textual substance. To this end, it may be appropriate to compare P2 with Pii again—the pair of representations of God enthroned—in an attempt to bring to the fore P2’s latent meanings through structural opposition to Pii. The two pictures constitute a variation on the motif of God enthroned, as stated above. On the one hand, Pii is a vision of the Deity seated on the throne in the celestial temple, flanked by seraphim and cherubim; as such, it is a visualization of the perfect unity and harmony of heavenly existence. On the other hand, P2 depicts the enthroned Deity in relation to his closest subjects, with a focus on Lucifer. At issue here is the irreversible hierarchical relation between the Creator and Lucifer, rather than a particular meeting scene (as claimed by Ohlgren 1972b, for example; see above), whether submissive or challenging. From such an abstract relational perspective, the Lord is depicted on the periphery and with lesser detail than in Pii, because, in contrast to Pii, the primary concern of the P2 drawing is not with the Deity himself but in relation to other celestial creatures. And yet God is represented by far the largest of all figures in the picture on account of his absolute status. In marked contrast, while occupying the center, Lucifer is portrayed smallest of all as a child. Of paramount interest, the two parameters, the small–large opposition and the center–periphery counterpart, are effectively counterbalanced with the covariation of small/center and large/periphery, so that the two figures involved—God and Lucifer—may be perceived equipollently as major focal figures in the scene. Either character is assigned a positive value—center or

32 

 2 God Enthroned: Pii/P2, with a brief discussion on P3.2b, P16.1, and P17.1

larger—for one of the variables at hand, so that neither receives overwhelming prominence at the expense of the other. By depicting Lucifer as a child, not just being relatively smaller in size and standing on the lower ground level than the Creator, the artist would have intended to emphasize most eloquently Lucifer’s inherent dependence on and subordination to the Deity. Inasmuch as P2 may be conceptualized as a preamble to the episode of the rebellion and fall of angels, Lucifer must appear in the center, as in the first two panels in P3; Lucifer, whether represented as a small child or a powerful leader, must equally be aligned with the central axis in the three consecutive panels. In visualizing the hierarchical relation between God and Lucifer, presumably the primary content of the drawing, the latter is represented as a youngster for the purpose of highlighting the absolute difference in status within the existential hierarchy between the two beings. Such a technique thus may be understood as a rhetorical device on the pictorial plane, a visual hyperbole, to be more specific. Incidentally, the artist made additional use of a halo to guarantee that viewers would correctly recognize his intended identity of the child as Lucifer, the angel highest in heavenly ranking next to God. Given that the relation between God the Creator and Lucifer the created is inherently asymmetrical, the entire constellation of figures in P2 cannot be shaped in any other way. In this light, the asymmetrical configuration as a whole of God and his subjects in P2 makes perfect sense as an iconic sign of the fundamental differences between the two parties involved (one vs. many; Creator vs. created; self-existing vs. dependent), and thus stands in polar opposition to the symmetrical figuring of the Lord attended by seraphim and cherubim—the consolidated body of the heavenly ruler—depicted in Pii. It should be noted that the representational asymmetry derives inevitably from the perspective taken here: consequently, represented in three-quarter view, even the Deity, who is otherwise expected on intrinsic grounds to be depicted with full symmetry as in Pii, is composed asymmetrically. The total absence of symmetry at the multiple levels of visual representation in P2 may therefore be construed as signifying at large the immanent disharmony in, and the imminent upheaval of, the heavenly order symbolized in Pii. Furthermore, since the Creator and the created are separate in their existence, the figures in P2 are each presented in discrete and individuated form. By contrast, the Deity conceptualized as a self-existing absolute being, the Godhead, or the Creator before Creation (cf. Karkov 2001: 48) is represented as all-encompassing, with a consequence that his attendants have inseparably been incorporated in his expanse. Compare in this connection Isaiah 66.1: “Thus says the Lord: Heaven is my throne and the earth is my footstool . . .” (Coogan 2010). Thus, the physical autonomy in P2 is diametrically opposed to the disembodied representation in Pii.28 28 Further questions remain to be answered, however. Why is the Deity represented in the right corner, rather than in the left? More generally posed, why is the grouping of figures crowded toward the right periphery? Why are a larger number of figures represented to the left of the Deity, rather than the other way around? Principled answers to these questions will be presented in section 2.2 below.

2.1 A synchronic–structural perspective 

 33

The opposition between symmetry and asymmetry treated above also bears on the way a frame is constructed. As may be recalled (Table 3), Pii is provided with a symmetrical and complete framing, whereas P2 is enclosed in an asymmetrical and incomplete manner. The analogous alignment of symmetrical/perfect with Pii and asymmetrical/ imperfect with P2 in figure representation offers a unitary overall account of the correlation in framing. Yet beyond this general understanding, more specific explanations may well be explored. As discussed above, the setting of the scene in the temple is closely associated with a vision of God enthroned, such as is represented in Pii. Being a variant of the same motif, however, the primary interest of P2 is shifted to God enthroned conceived not on his own, but in relation to Lucifer, and derivatively to his other attending angels. With this shifted focus onto the hierarchical relation between the Creator and the created, the significance of the original association with the temple would have been attenuated, giving way to the foregrounding of individual figures. Accordingly, the architectural frame with arch and capitals would have been dispensed with, thereby motivating the minimal depiction of framing lines. It may be interesting to point out in this regard that removing the arch from Pii (as motivated above) and connecting the left and right tops of the frame lines with a comparable double line (as a consequence of minimalizing the frame)—that is, the flattening of the arch—forms crossings with the upper ends of the seraphim’s wings, configurations exactly obtained in P2. Moreover, the considerably shorter left border in P2 may be viewed as a trace of the column with a capital; more specifically, the existing short border on the left is a reflection of the original part above the capital, with the lower part completely removed. This may give a credible account of why the left border does not reach the ground level, the otherwise expected extension, but stays at a point considerably higher—one of the questions posed above and left unanswered so far. Since the original column was structurally divided into three segments (the capital, the shaft, and the base), the place corresponding to below the capital naturally would have served as a reference point for trimming.29 The conceptualization of the Creator as an all-encompassing self-existing being in Pii in contrast to the Creator–created axis of opposition in P2 may explain the polarization between the two pictures in regard to the attendants’ orientation (Table 3). Merged into a unified, symmetrical being, all the constituent faces in Pii look outward in perfect harmony with the Creator’s own outward look. Correspondingly, the two seraphim, lacking hands and feet of their own, are partly fused together (at least not represented in an explicitly separate form) as integral with the Creator. By contrast, the hierarchical polarity between, and the concomitant individuality of, the Creator and the created come to the fore in P2. Therefore, while the attending seraphim, separated from the Deity as well as from each other (no overlap), are fully and individually depicted, they face toward the Lord as the source of their blessing and target of glorification.

29 By contrast, the right border is retained on functional grounds, as discussed above.

34 

 2 God Enthroned: Pii/P2, with a brief discussion on P3.2b, P16.1, and P17.1

Since the reading presented above for P2 differs significantly from others conventionally held, it is appropriate to make explicit the text–picture alignment that it entails in distinction from earlier interpretations bearing on, for example, Genesis A 22a–28a/ Genesis B 259a–264b, 292a–295a (Ohlgren 1975: 39–41), or Genesis A 31b–36a/Genesis B 259a–271b (Broderick 1978: 73). As far as Genesis A is concerned, it is verses 12b–21b that constitute a verbal equivalent of P2: hæfdon glēam and drēam // and heora ordfruman / engla þrēatas, // beorhte blisse. / wæs heora blǣd micel. // þegnas þrymfæste / þēoden heredon, // sægdon lustum lōf, / heora liffrēan // dēmdon: drihtenes / dugeðum wǣron // swīðe gesǣlige. / synna ne cūþon, // firena fremman / ac hīe on frīðe lifdon // ēce mid heora aldor. / elles ne ongunnon // rǣran on roderum / nymþe riht and sōþ (The hosts of angels had rejoicing and happiness, bright bliss and their source of being. Their glory was great! Triumphant attendants glorified the prince, eagerly spoke praise, honored their Lord of life, were very happy in blessings of the Lord. They did not know of sins, the doing of evil deeds, but rather they lived in peace, eternally with their leader. They strove to exalt nothing else but right and truth, …) (Anlezark’s [2011: 3] translation)

In contrast to the preceding passage (1a–12a), which is exclusively concerned with God and accordingly ties in optimally with the frontispiece illustration (Pii), at issue here are angels’ praise and glorification of God, and their resulting bliss and happiness.30 That is, the attendants are totally submitted to and dependent upon the Creator as their source of blissful being, fully in accordance with the heavenly order. As pointed out repeatedly in earlier scholarship, however, the illustration finds a better fit in Genesis B primarily because of the explicit mention of the brightest angel, most beloved of God, although unnamed in the text (e.g., Ohlgren 1975: 58–59; Raw 1976: 146–147). The most relevant passage is as follows (Genesis B 246a–258b): Hæfde se alwalda / engelcynna // þurh handmægen, / hālig drihten, // tēne getrimede / þǣm hē getruwode wel // þæt hīe his giongorscipe / fyligan wolden, // wyrcean his willan / forþon hē him

30 According to Broderick (1978: 56, 71), a longer portion ranging from 1a through 21b counts as a verbalization of Pii. By contrast, Karkov (2001: 45–46) singles out a shorter passage, Genesis A 1a–8a, as an accompanying text of Pii. With an exclusive reference to the Creator as an absolute being, however, the picture seems to match 1a–12a more closely, as follows: Ūs is riht micel / ðæt wē rodera weard, // wereda wuldorcining, / wordum herigen, // mōdum lufien. / hē is mægna spēd, // hēafod ealra / hēahgesceafta, // frēa ælmihtig. / næs him fruma ǣfre, // ōr geworden / ne nū ende cymþ // ēcean drihtnes / ac hē bið ā rīce // ofer heofenstōlas / hēagum þrymmum, // sōðfæst and swīðfeorm, / sweglbōsmas hēold. // þā wǣron gesette / wīde and sīde // þurh geweald godes / wuldres bearnum, // gāsta weardum. (It is very right for us that we should praise with words the guardian of the heavens, the glorious king of hosts, should love him in our minds. He is abundant in powers, head of all lofty creatures, the Lord almighty. There never was a beginning for him, nor an origin brought about, nor presently will come an end of the eternal Lord, but forever he will be sovereign over the thrones of heaven. Righteous and potent in supernal powers, he has held the expanses of heaven, which were established broad and wide through God’s rule for the sons of glory, for the guardians of spirits.) (Anlezark’s [2011: 3] translation)

2.1 A synchronic–structural perspective 

 35

gewit forgeaf // and mid his handum gescēop, / hālig drihten. // gesett hæfde hē hīe swā gesǣliglīce, / ǣnne hæfde hē swā swīðne geworhtne // swā mihtigne on his mōdgeþōhte, / hē lēt hine swā micles wealdan, // hehstne tō him on heofona rīce, / hæfde hē hine swā hwītne geworhtne, // swā wynlīc wæs his wæstm on heofonum: / þæt him cōm from weroda drihtne. // gelīc wæs hē þām leohtum stēorrum. / lōf sceolde he drihtnes wyrcean, // dȳran sceolde hē his drēamas on heofonum / and sceolde his drihtne þancian  // þæs lēanes þe hē him on þām lēohte gescerede  / þonne lǣte hē his hine lange wealdan. (The ruler of all, the holy Lord, had arrayed ten orders of angels by the power of his hand, whom he well trusted would follow in his obedience, work his will—because he granted them intelligence. And the holy Lord created them with his hands, he had established them so blessedly; one he had made so potent, so mighty in his intellect, so much he let him rule, highest after him in the kingdom of the heavens, so radiant had he made him, so delightful in the heavens was his stature that came to him from the Lord of hosts, that he was like the dazzling stars. He should have performed the Lord’s praise, should have cherished his joys in the heavens, and should have thanked his Lord for the reward that he bestowed on him in that light—then he would have allowed him to rule for a long time.) (Anlezark’s [2011: 19, 21] translation)

In parallel to the two variant pictures Pii and P2 discussed above, there is a comparable trio that varies along a subset of the same parameters that differentiate Pii from P2. At issue are the latter (or lower) half of the second register in P3 (Pl. 3) and the upper representations in P16 (Pl. 10) and P17 (Pl. 11), henceforward referred to as P3.2b, P16.1, and P17.1, respectively. In addition to their inherent compositional and iconographic commonalities as discussed immediately below, these three pictures are exclusively followed underneath by the scene of the falling/fallen angels (P3.3, P16.2, and P17.2). In this light, it may be legitimate to treat the three pictures as a unified group. According to my new interpretation of P3.2b, to be substantiated at length in section 3.1 below, God, seconded by two angels on the right, points javelins at Lucifer and his company on the left. In P16.1, God, flanked by three angels on either side, looks to the left, addressing the angels standing on that side. According to Gollancz (1927: xli), these angels on the left side are rebellious ones. In support of his interpretation, one may note that the angel to the immediate left of God wears a crown and holds a palm in P16.1, a pair of attributes reminiscent of God’s major antagonists depicted in P3.1 (the central figure with a crown and a scepter), P3.2a (the central figure with a crown and palms), and P3.2b (the figure with a crown, standing to the left of God). Finally, P17.1 represents God enthroned set within a mandorla and accompanied by seraphim. Of special interest are the particular ways in which these three pictures—P3.2b, P16.1, and P17.1—are contrasted, and their overall organization in relation to other relevant pictures. Specifically, P3.2b/P16.1 and P17.1 are distinguished by a complementary clustering of features that constitute a subset of the parameters identified in Table 3 above as responsible for the variation between P2 and Pii. Among the wealth of features distinguishing Pii from P2, seven properties are involved in the opposition between P3.2b/P16.1 and P17.1: (i) frame design; (ii) drawing space; (iii) depiction of God (or figures in general); (iv) God’s face orientation; (v) number of attendants; (vi) God’s posture; and (vii) configuration or arrangement of constituent figures (Table 4). While P3.2b/P16.1 are bound on both sides by simple columns, P17.1 is framed with elaborate

36 

 2 God Enthroned: Pii/P2, with a brief discussion on P3.2b, P16.1, and P17.1

towers. The space allocated for representation is relatively larger in P17.1 than in the other two. God and the angels are fully depicted in P17.1, whereas they are cropped with their feet not entirely represented in P3.2b/P16.1. God is presented full-faced in P17.1, but represented looking sideways in P3.2b/P16.1. God is flanked by a minimal number of attendants (one on either side) in P17.1, but more than one in P3.2b/P16.1. God is seated in P17.1, but standing in P3.2b/P16.1. God and his company are arranged symmetrically in P17.1, but asymmetrically in P3.2b/P16.1. The distribution of these features is accordingly summarized in Table 4. Table 4. Oppositional features of P3.2b, P16.1, and P17.1a Feature

Value for P17.1 (analogous to Pii)

Value for P3.2b/P16.1 (analogous to P2)

Frame design Drawing space Depiction of God (or figures in general) God’s face/body orientation Number of attendants Posture Configuration

Elaborate Larger Detailed

Simple Smaller Abridged (no feet)

Frontal Minimal Seated Symmetrical

Profile Multiple Standing Asymmetrical

a

To facilitate comparison with Table 3 above, P17.1 is presented to the left of P3.2b/P16.1.

No less significant than the features themselves is their actual alignment with each other. With the exception of posture (seated vs. standing) uniquely applicable to the triad of images under discussion, the six remaining features involved in the distinction between P17.1 and P3.2b/P16.1 cluster in much the same way as in the contrast between Pii and P2. That is, what Pii is to P2 is largely identical to what P17.1 is to P3.2b/P16.1. Since these properties are independent of each other, the exact pattern of their distribution cannot be attributed to pure contingency, as the probability of the same cooccurrence pattern obtained in any two pictures is less than 0.001. It may therefore follow as a matter of course that P17.1 and P3.2b/P16.1 should be grouped with and subsumed under Pii and P2, respectively, as their variant representations. Furthermore, since the latter two (Pii and P2) stand on their own as autonomous pictures whereas the former are constituents of larger illustrations, P17.1 and P3.2b/P16.1 may be characterized as simplified or miniature versions of Pii and P2, respectively. The five related pictures under consideration, Pii, P2, P3.2b, P16.1, and P17.1, then, are integrated into the system that is organized on the basis of two parameters, each ranging between two oppositional values. Generalizing over the specific concrete features that have been determined as characteristic of the pictures in question, as listed in Tables 3 and 4 above, one may postulate the two overarching parameters as follows:

2.2 A diachronic–comparative perspective 

 37

(i) the mode of representation (independent vs. dependent) and (ii) the mode of activity (inaction [or static] vs. action [or dynamic]).31 The first parameter concerns the status of the illustrations being executed: are they presented as standing on their own or as part of a larger picture? On this parameter, Pii and P2 are contrasted with P17.1, P3.2b, and P16.1 because of their autonomy as self-standing representations. The second parameter has to do with whether the central figure, God, is represented acting on another figure in a given picture; in other words, is God engaged in a transitive action, an action that affects another being? By this criterion, Pii and P17.1 are categorized as lacking such an action (i.e., inaction). In contrast, P2, P3.2b, and P16.1 involve action: God is addressing Lucifer (and his host) in P2 and P16.1, while he is threatening or attacking Lucifer and his company in P3.2b. In summary, the five variant pictures are organized into the following quadripartite system predicated on the two binary parameters. Table 5. The system of organization for variant pictures of God enthroned (Pii, P2, P3.2b, P16.1, and P17.1)  

Inaction

Action

Independent Dependent

Pii P17.1

P2 P3.2b, P16.1

2.2 A diachronic–comparative perspective The foregoing synchronic–structural examinations of Pii and P2, which have brought about their mutually illuminating and reinforcing insights and interpretations, draw on the working hypothesis that the two illustrations constitute a variation on the same motif, God enthroned. This assumption is a claim that is put in stronger and more substantial terms than a simple acknowledgement of superficial similarities between the two pictures. Specifically, it implies from a diachronic perspective that the two drawings were produced in essence after a single common model with different purposes in mind. Whatever the resultant dissimilarities attested in the two existing representations would have arisen from varying adaptations and alterations, intentionally made for the most part but at times accidentally committed, that were introduced in the course of (re) production and transmission. In this section, I will explore the major diachronic–genealogical implications of this hypothesis. Since the underlying motivation for production is not a faithful copying of the model, it naturally may be surmised that numerous features of the original underwent alteration including total elimination, and, conversely, that novel traits and interpretations were introduced by innovation and reanalysis. 31 On Schapiro’s (1973) pioneering work on the opposition between theme of state and theme of action, see section 12.2 below.

38 

 2 God Enthroned: Pii/P2, with a brief discussion on P3.2b, P16.1, and P17.1

At this point, following the helpful lead of historical–comparative disciplines like linguistics, the problematics may be reformulated as follows: Pii and P2 are variants of the same earlier form; they alternate with each other under specific conditions in a given synchronic state (Junius 11), as investigated in the preceding section. The synchronic alternation thus established was a consequence of a complex of changes that occurred on the originally single item (their common model). The purpose of historical– comparative investigations, then, is to determine these intervening changes, identify their causes and the mechanism for their occurrence, and ultimately recover the original state of affairs before the split into the alternants. Obviously, the metaphor is a gross simplification, as art historical problems are far more complicated than linguistic ones: while the latter are largely mechanical in causation and often beyond human intentions, the former are deeply implicated in complex cognitive and emotional activities. Given that the two drawings are derived from the identical model, and given further that these two variants are differentiated through an alternative choice of binary values for the set of distinctive features listed in Table 3, one may ask the following question in reconstructing the model: Which of the two reflex pictures inherits which attributes from the source? In practicing reconstruction on the basis of comparative data regardless of disciplines, one is usually guided and constrained by a general probabilistic tendency for devolution, degeneration, and simplification: given two varying values for a property, the more specific, elaborate, complex one is more likely to reflect the original. Put another way, the attributes that require greater expenditure for realization are more likely to be attributable to the common ancestor. This guideline allows for exception and violation, to be sure, but such exceptional cases require special pleading and independent substantiation. In accordance with the above practicing guide, one may reconstruct the value set for the nine oppositional features (features 1 through 9, as shown in Table 6). Fully providing a symmetrical frame in architectural form, and using comprehensively the drawing space primarily for a detailed and prominent depiction of a focal figure will no doubt take a greater amount of preparation and execution than illustrating a central figure small and leaving vacant a large part of the available space that is incompletely enclosed by a simple borderline. Since the former clustering of feature values (1 through 6 in the following table) distinguishes Pii from P2, one may be led to infer that Pii is more faithful to the model in these respects. For the remaining three features, however, the drawing of P2 must be characterized as following the original more closely. By displaying two additional angels beside the seraphim (feature 7), P2 supersedes Pii in representational intricacy. Much the same can be said of the extent of individuation of figures (feature 9): depicting individual bodies discretely including their limbs, the illustration in P2 should be rated higher than the continuous and cohesive configuring of otherwise distinct bodies in Pii. Further justification for inferring the composition of Pii as nonoriginal and hence innovative is the apparent uniqueness of such a conglomeration of the Deity, seraphim, and cherubim (cf. Broderick 1978: 58). Statistically speaking, such a singular property is less likely to recur than occur just once for whatever reason,

 39

2.2 A diachronic–comparative perspective 

unless a comparable form is attested independently and a cogent case can be made for their genealogical relatedness. One also must call attention to feature 8 with respect to the compositional uniqueness of Pii. In the complete absence of a representation in which flanking attendants all look outward away from their ruler, it is more reasonable to assume that the model would have manifested the normal pattern of inward-facing orientation, particularly where a group of seraphim larger than a pair is at stake. To be added to the inventory of contrastive features that are found in complementary distribution between Pii and P2 and accordingly may be reconstructed for their common model is the outstanding anomaly shared by the two drawings, that is, the absence of a mandorla (feature 10 in Table 6). In view of the almost invariable enclosure of the enthroned Deity by a mandorla or a circle, its nonoccurrence in the two pictures is more likely to have been inherited from their common source than to have been designed each by independent innovation. Drawing on the inventory of features thus reconstructed, let us search for a probable model that most likely would have served as a primary source for Pii and P2. Needless to say, the above features constitute necessary, but not sufficient, conditions for identifying the source. In determining the model picture firmly, one must also pay close attention to other nonoppositional features, that is, fine details in style, composition, and design elements that have evaded structural examination up to this point. Specifically, the identification of a particular picture (or a set of pictures) as a plausible source of inspiration for Pii and P2 would be confirmed and reinforced on empirical grounds to the extent that a number of otherwise unintelligible and unexplainable singular traits become susceptible to revealing accounts by ascribing them to specific analogs in the alleged model as their sources of inspiration and explaining the mechanisms of and motivations for their adaptations and derivations. Table 6. Reconstructing feature values for the common model based on their reflections in Pii and P2  

Feature

Value for the model (= Vivian 423r)

Pii

P2

1  2  3  4  5  6  7  8  9  10

Frame shape Frame design Framing Drawing space Figure depicted at the center Depiction of God Group of attendants Orientation of attendants Contour of figures Mandorla

Symmetrical Elaborate (architectural) Complete Fully occupied Largest being (the ruler) Detailed More than minimal Looking inward (toward the ruler) Discrete; individuated Absent

√  √  √  √  √  √        √ 

            √  √  √  √ 

40 

 2 God Enthroned: Pii/P2, with a brief discussion on P3.2b, P16.1, and P17.1

In searching for the most probable model, Pii and P2 in all likelihood would have been inspired (in their own different ways, as will be elucidated below) by the Vivian Bible or the First Bible of Charles the Bald (Paris, Bibliothèque Nationale de France, Ms. lat. 1), fol. 423r (Pl. 32; Köhler 1933: 60–64, 220–231; Kessler 1977: 125–138; Schramm 1983: 51–53; Diebold 1990: 85–133; Dutton/Kessler 1997: 71–87; Garipzanov 2008: 244–247). The picture in question appears at the end of the volume as a presentation miniature, representing the scene in which the Bible is being presented to Charles the Bald by the monastery of Saint-Martin in Tours in 845. The presentation ceremony is set in the large space enclosed by the elaborately designed, symmetrically constructed architectural frame comprising a draped arch and spiral columns with acanthus capitals and bases (features 1, 2, 3, and 4; Table 6). Charles the Bald is depicted in the center, the most prominent figure in the scene, seated on the throne, holding the scepter in his left hand, and raising his right, pointing in the direction of the Bible about to be presented (features 5 and 6). At either side and somewhat diagonally behind, a retainer flanks the king, grasping with one hand the outer frame of the back of the throne. Next to these two attendants but slightly advanced forward stand two bodyguards on duty. In front of the king on the lower level, a group of monks and a layman (nontonsured), who must be Vivian himself, assemble in a semicircular formation, in which the people on either side face toward those on the opposite, all raising their hands in front—with a possible exception of one person who seemingly looks up toward the king; among these monks, two immediately in front of the guard on the left side hold the Bible (= the Vivian Bible) together, and are about to present it to the king. Thus, at first glance, an assembly of sixteen people (two retainers, two guards, two monks bearing the Bible, nine other monks, and a layman standing opposite the Bible bearers), each facing inward and depicted with individuality in appearance and posture, encircles the secular ruler, who in his turn is not portrayed with a mandorla or the like (features 7, 8, 9, and 10). On closer examination, however, the four monks on either side of the one standing direct opposite the king and turning his back toward the viewer should be taken to be of the same identity to each other (Dutton/Kessler 1997: 75–78). More specifically, the group is situated on the left side before the presentation of the book, and then changes places to the right side, turned around, and facing the monk at the bottom center, forming a two-by-two square assembly. Vivian 423r is thus inherently dynamic in representation: rather than a monolithic moment’s view, it provides a condensation of a series of acts extending for the duration of the presentation ceremony. The book is presented from the left by the two monks to the king, who, then repositioned to full face, presumably places it on his left knee in accordance with the motif of God enthroned; meanwhile the group of four monks processes past the central monk in the bottom layer to his right and, led by this man, sing psalms to the king in praise; thereupon the king gazes in that direction in appreciation. As far as the face orientation of the king is concerned, it comprises three successive scenes: from three-quarter view facing left to frontal (analogous to the figure of Charles the Bald in San Paolo 1r; Cardinali 2009: 19) to three-quarter view facing right (similar

2.2 A diachronic–comparative perspective 

 41

to the portrait of Lothair in the Gospels of Lothair, Paris, Bibliothèque Nationale, Ms. lat. 266, fol. 1v; Mütherich/Gaehde 1976: pl. 25). As for the location of the Bible, it is first held by the two monks, and upon presentation it likely rests on the knee of the king under the left hand. In this light, Pii and P2 seem to capture the first two phases of the king’s acts/ states in terms of his face orientation. As substantiated in the foregoing account, Vivian 423r embodies all of the structural features that have been reconstructed through a comparative examination of Pii and P2 for their common model. Since these characteristics are merely a small subset of the whole inventory of features constitutive of the pictures at issue, one must reformulate one’s problematics in scope and depth by going beyond identifying structural parallels as has been done above, and addressing squarely the motivations for and mechanism of adopting and adapting Vivian 423r for the drawings Pii and P2. Particularly at issue is the assemblage of singularities of the latter two in their commonality to and deviation from their comparable components in the source picture. Let us begin with P2, which seems to be more conspicuously similar in a way to Vivian 423r in terms of figuration and composition, as illustrated in Fig. 1. The primary concern of P2 is to represent the heavenly hierarchy in which Lucifer is to be totally submissive to the Creator, as discussed in the preceding section. From this perspective, the artist would have reorganized the model picture of Vivian 423r to best suit his purpose. While Charles the Bald is optimally reused as God, it has yet to be determined who should be chosen as Lucifer among the remaining figures, who are all the king’s subjects. At this point, the king’s gaze may serve as a principal criterion for adaptation. Since Charles looks toward the Bible to be presented (cf. Köhler 1933: 224), Lucifer should be placed in that direction. In the accompanying poetic text (Genesis A), however, Lucifer features as a central figure (or a topic to use a text-oriented notion). Accordingly, it is most natural that Lucifer should occupy the center of the drawing space, rather than the margin as in the model representation. Such a centralization of Lucifer on the horizontal axis, implemented in concert with his comparably focal position in the first two registers on the immediately following page (P3), that is, P3.1 and P3.2a (see section 3.1), accounts for the resultant asymmetrical arrangement of figures in P2, in contrast to their original more or less symmetrical configuration in Vivian 423r. More specifically, while creating an open space on the left margin, the overall shift to the right now relocates the Creator from the center to the right periphery, therewith displacing the original retainer in part—now redrawn as the right seraph—and the bodyguard in full, along with Vivian immediately below.32 While the latter two—the bodyguard and Vivian—get completely trimmed off, the former— the retainer—has a lesser space available than can be fully represented. Consequently,

32 Taking into account the treatment of the sword and the curtains, it may be more appropriate to postulate a conflation of the two men into the right seraph, with the consequence that the two objects contiguous to the bodyguard left their traces in the recomposition as P2.

42 

 2 God Enthroned: Pii/P2, with a brief discussion on P3.2b, P16.1, and P17.1

the seraph on the Lord’s right is pictured limbless and has only two of the four lower wings depicted, in contrast to the seraph standing on the other side, who is properly equipped with limbs and wings (section 2.1). While seraphim display the spread palms alone, leaving the remaining part of the hands covered with wings—as exemplified in P17 (Pl. 11) and San Paolo 117r (Pl. 35)—the depiction of larger portions of the hands above the wrist in P2 may be attributed to the underlying attendants of Charles the Bald in Vivian 423r. Moreover, the posture of the right hand of the seraph standing to God’s left is almost identical to that of the corresponding retainer in the Vivian scene. Furthermore, the conventional figuration of a seraph in symmetrical shape—as exemplified in Vivian 130v (Pl.  28), San Paolo 117r (Pl.  35), and Boethius 158v (Pl.  73)—may account for why not only the leftmost wing of the right seraph—due to lack of space—but also its rightmost counterpart—despite the narrow strip of space otherwise available for its fragmentary representation—are removed altogether.

Figure 1. Figural and compositional correspondences between Vivian 423r (adapted, left) and Junius P2 (detail, right). Paris, Bibliothèque Nationale, MS lat. 1 (Vivian Bible), fol. 423r (detail, adapted), Dedication frontispiece. Source gallica.bnf.fr/Bibliothèque Nationale de France (licensed under CC BY 4.0). Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 2, God Enthroned, Facing Lucifer (detail). Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).

On the vertical axis, given that the Lord–Lucifer line of sight is of overriding importance, other figures standing beneath that level go out of view, thereby leading to their maximal obscuration in representation—almost total disappearance from the space below the wavy ground line in P2 (Pl. 2a; see further below). As distinctively demarcated in different colors, the presentation miniature (Vivian 423r, Pl.  32) consists of three different layers for setting figures, the upper (white), the middle (blue), and the lower (green) level. In this light, it may be generalized that, as far as the vertical axis is concerned, the top level was exclusively adopted for recomposition, because it stands

2.2 A diachronic–comparative perspective 

 43

closest to the celestial realm (with which the artist is concerned), as symbolized by the cloudy ground that seemingly extends up to heaven without disruption. Consequently, the Bible (and by extension its bearers) belonging to the intermediate level, which contradictorily would have had to have been left undepicted, was slightly relocated upward into the top sphere; correspondingly, the leftmost angel—derived, in terms of location relative to the king, from the bodyguard carrying a lance and a shield—was raised to a higher position than its Vivian preform.33 Thus, the particular reductive ways in which adaptations were implemented were determined structurally through a negotiation between the three-layered organization inherent to the original picture and the new artistic purposes and perspective for recomposition. In addition to the eclipsing effects consequential to the focusing of the Lord– Lucifer axis, the total elimination of a semicircular assembly of figures would have been induced by another motivation. The large blank space below the ground would have been designed as an icon for the nonexistence of hell, which had yet to be created for banishment and torture of traitors (Genesis A 36b–38b), as vividly illustrated on the facing page (P3). Ultimately, the open space may signify that “nothing has come about yet here,” as stated in Genesis A 103a–106a: “Ne wæs hēr þā giet / nymþe heolstersceado // wiht geworden / ac þes wīda grund // stōd dēop and dim / drihtne fremde, // īdel and unnȳt” (Then was nothing yet here except darkness, but this vast abyss stood deep and dark, alien to the Lord, idle and useless; Anlezark’s [2011: 9] translation). In this way, far from being an “improper placement” or “[i]nexpert copying” (Ohlgren 1975: 41), the apparently off-balance arrangement of figures and their diverse individual representations would have been the artist’s well-designed adaptation of the underlying model to best suit his purposes. Similarly defocused, the elaborate architectural frame loses its original significance and is accordingly simplified into a minimal double line in P2, in diametric contrast to Pii, in which the arch and columns with capitals are partially retained. The extended double line on the right may be viewed as confirmation that the lower space, currently blank (P2) but originally pictured (Vivian 423r), constituted integral part of the model picture, as is actually true of Vivian 423r. As far as the depiction of the Lord is concerned, particularly noteworthy are three points of detail, all involving adaptation of the original scepter in its shape and its positioning. First is the transformation and transference of the original scepter. In Vivian 423r, a long scepter—a relatively long rod extending from the shoulder at the top to the footstool at the bottom—is grasped in its middle section by the king with the left hand, which rests on the left lap. In P2, by contrast, a much shorter but comparatively thick stick—referred to as a scroll/scepter, for lack of a definitive categorization, in the foregoing discussion—is placed on the left lap at the bottom end, held with the left hand in its

33 In terms of posture and gesture, however, the angel in question seems to reflect one of the monks, particularly the one standing to the immediate left of the conducting figure, as illustrated in Fig. 1 (left).

44 

 2 God Enthroned: Pii/P2, with a brief discussion on P3.2b, P16.1, and P17.1

middle, and extends diagonally to his right up to a point that is far away from the shoulder in height and distance, and close to the seraph standing on God’s right side. This new shape and positioning of the scroll/scepter relative to the Lord proves strikingly similar in compositional terms to that of the rear post of the back of the throne in Vivian 423r, which is gripped by the retainer attending on the king’s right. Also not to be left unmentioned is the sword that is held by the rightmost guard (Pl. 32). The relative thickness and length, if not the location in relation to the king, would seem to be reminiscent of the scroll/scepter in P2. These three rod-shaped items—all grasped in the hand—in the model picture would have served as a perceptual basis of analogical recomposition for the artist. In any event, in P2, these three rods are apparently merged into one (with the exception of the clavus left behind as a trace of the former scepter; see below), whereby the resultant scroll/scepter is transformed to be shorter and thicker than the original scepter, and positioned at a distance from the torso, with the bottom end resting on the left lap and the top located beside the attendant’s head. Second and inseparable from these alterations is the vertical position of the king’s and the Lord’s left leg. While it is approximately the same level in height as the right leg in Vivian 423r, it is raised considerably higher in P2. Therefore, one may conjecture that in adapting Vivian 423r for P2 the artist would have reconfigured—for reasons yet to be clarified—the scepter and other major constituents of the left side of the body relative to each other as well as to some of those lying outside farther away. Finally, the clavus on the Lord’s left torso in contrast to its appearance on the right side in Pii should be brought to bear on the adaptive reconfiguration at issue. Of particular interest, the Vivian model does not depict a clavus. Instead, the king holds a scepter from his left shoulder down to the footrest in the model. Normally, however, the clavus is visible on the right side, as exemplified by the vast majority of representations of God enthroned. Therefore, it is unlikely that the appearance on the left side in P2 would have been spontaneous. In more specific terms, the artist would have been led to reinterpret the upper portion of the original scepter as a clavus, in consequence of the reconfiguring of the left hand, the scroll/scepter, and the left knee (as suggested above), which had made it impossible to retain the earlier scepter without changing its original identity as such. While the compositional parallels thus have been identified between Vivian 423r and P2 and the overall trajectory of derivation from the former to the latter can be outlined with reasonable confidence, it remains to be explained why these specific alterations were implemented as adaptive tactics and how they were related to each other if they really were. Of paramount importance here are therefore the questions of underlying motivations for choosing the set of concrete adaptations as they were actually executed, and of causal and functional interrelations among these multiple reworking practices. Choosing Vivian 423r as a primary model for new compositions of God enthroned— Pii and P2—in the Old English/Old Saxon Genesis as recorded in Junius 11 (or more plausibly its Old Saxon antecedent; section 12.8), the artist also would have been guided by

2.2 A diachronic–comparative perspective 

 45

another image, namely, the exemplary representation of Christ in Majesty in works of the School of Tours (as exemplified in the Vivian and San Paolo Bibles):34 the Creator with a frontal face is seated on the throne, holding the book with the left hand on the top and with the left lap underneath, while the left leg is positioned higher than the right (cf. Köhler 1933: 240). As will be shown below, this typical representation is embodied with minimal modifications of adding a scroll/scepter in Pii. As for P2, however, this conventional Lord image conflicts irreconcilably with the three-quarter view that constitutes the grounding perspective for the P2 representation: this drawing is intended to be about the Lord–Lucifer relation, for which the line of sight must be determined exactly as in the model (423r) as opposed to the exemplary frontal image of the enthroned Deity. Moreover, the presence of the book on the knee characteristic of the convention loses significance here: in the underlying presentation miniature, Vivian 423r, the book is still in the hands of monks, awaiting presentation to the king, who gazes at the gift at a distance. Thus, of the two redrawings (Pii and P2), Pii constitutes the more transparent reflection of the exemplar of Christ in Majesty. In comparison, inspired all the same but affected less radically by the Touronian Christ image, P2 remained correspondingly closer to the Vivian model (423r) as far as the depiction of the Lord is concerned. Without any need to change face orientation and depict the book on the knee in conformity to the exemplary image of God enthroned, all the artist had to do was, first, raise the left leg, and, second, recategorize, reshape, and reposition the scepter—now the scroll/scepter. While originally necessitated for Pii, these alterations were found also compatible with the defining design of P2. Consequently, this picture was analogously adapted, presumably for the purpose of maximizing the compositional correspondence between the pair of variants at issue. These two adaptations executed for P2 seem linked in operation, whereby the raising of the left leg would have triggered the reconfiguration/recategorization of the scepter originally represented in Vivian 423r. In precise terms, the Lord’s left leg in the exemplar is positioned not only vertically higher than the king’s in Vivian 423r, but also diagonally rightward (from the viewer’s perspective), as seen in 329v (Pl.  30) of the same manuscript. Given that the scepter is appended to the left knee in 423r, the upper rightward raising of the left leg would naturally incur the relocation of the attached scepter in the same direction, resulting in a reconfiguration whereby the upper portion of the scepter virtually overlaps the rear post or stile of the throne that is grasped by the retainer standing to the king’s right. Moreover, since the stile in question structurally does not extend to the bottom but terminate at the seat, the resultant rod-like item, the former scepter, is shortened by half. Thus, the upper rightward shift of the left leg would have led to the analogous relocation of the scepter, which in turn would have been reshaped analogous with the stile of the throne that it superimposed, including

34 For details on the origins and evolution of the images of Christ in Majesty in the Touronian Gospels and Bibles, see Köhler (1933: 132–136, 240–245) and Kessler (1977: 36–58).

46 

 2 God Enthroned: Pii/P2, with a brief discussion on P3.2b, P16.1, and P17.1

the grasping hand. Also partly influenced by the sword held by the guard—another configuration of a rod and a hand—the original scepter ended up being a shorter, thicker rod, gripped by the Creator, the composition that is found at first glance in P2. The exact identity of this item, however, hardly succumbs to determination, as many accounts still compete with each other (cf. Broderick 1978: 56n2). After all, one may well view the item as inherently polysemous, open to multivalent, mutually enriching—rather than excluding—interpretations at the same time. Furthermore, as pointed out in the preceding section, this enigmatic object also would appear to be grasped by the seraph, the reading that is prompted not least by the feathers apparently extending from the roots of the fingers gripping the scroll/scepter. This supplementary reading may prove to be yet more credible in light of the underlying model picture: the stile of the throne is originally grasped by the retainer, rather than the king; and the original sword is held by the bodyguard. Under this alternative reading, the Lord’s left hand, unillustrated as it is, must be located on the left knee, as explicitly depicted in Vivian 423r. After the relocation, the original scepter did not disappear from its old location completely. It left behind a trace in the new composition: as discussed above, the clavus on the left side of the Lord’s body—contrary to the usual location—may justifiably be regarded as a reinterpretation of the scepter in the new setting. In conclusion, the raising of the left leg, the reconfiguration and recategorization of the scepter, and the transposition of the clavus in all probability would have constituted a chain reaction, each earlier event causing the next in sequence. The assumption underlying the foregoing discussion is that the Creator’s left leg was level with the right in the source material (Vivian 423r), and that the raised position of it in the current drawing was due to compliance with the exemplary image of the enthroned Deity prevalent in the School of Tours. Since the exemplar is the ideal conceived in the abstract, independent of individual realizations that are subject to variable contingencies in detail, it does not dictate particular ways its optional features are to be realized in individual productions. From this perspective, the noninherent, secondary status of the raised left leg gains plausibility with reference to the peculiar way it is depicted in P2 and also Pii (on this drawing, see further below). In these Junius 11 pictures, the raised left leg does not rest on something firm, but stays hanging in the air in such a manner that it looks as if, as Broderick (1978: 74) put it in regard to P2, “The Creator figure is seated precariously on a backless throne . . .” By contrast, in the exemplary representation of the enthroned Lord the left foot is usually placed upon a base correspondingly slanted so that the left side is raised higher. This marked difference in the grounding of the left foot may be explained by assuming that the raising of the left leg was not integral to the original model (Vivian 423r), but supplied secondarily in abstract shape from the exemplar, in which the grounding of the leg at issue is not specified at length, as it varies widely in individual instantiations. Had the artist followed a particular example, he would have copied the model’s own way of grounding the left leg that was manifested in that specific representation: accordingly, the foot would have been grounded firmly on something solid.

2.2 A diachronic–comparative perspective 

 47

Drawing on the claim substantiated above that the raising of the left leg laid the foundation for the reworking of Vivian 423r as P2, one may analogously explore the line of development from the same source to Pii. Again in conformity with the exemplary image of God enthroned, the level positioning of the two legs is modified so that the left is raised higher than the right and shifted toward the upper right. Furthermore, following the typical representation, the Bible is placed on the lap, which in turn is held with the left hand on the top, as exemplified in Vivian 329v (Pl. 30). The artist then is confronted with a challenge whereby he has to integrate the conflicting images into a coherent whole: on the one hand, he has the exemplary configuration of the raised left knee, the book, and the left hand; on the other, he has the three-quarter view of the king holding a scepter with the left hand that is placed on the left knee. Since, in contrast to P2, Pii is intended to represent the enthroned Lord as an absolute heavenly being, the three-quarter view does not justify retention, but must yield to the frontal one as an optimal image. As may be readily recalled, Vivian 423r constitutes a condensation of pictures moving in short succession, running from the moment immediately before the presentation of the Bible to the event after the presentation. In this light, the frontal view inheres in the model, despite first appearances to the contrary. In other words, the 423r picture is inherently capable of accommodating an explicit representation of the ruler’s frontal image in compositional and cognitive terms. The recomposition of the original three-quarter view as frontal therefore may be characterized as a natural development of the model, true to its inner logic, and far from being forcefully and arbitrarily executed from the outside. The implicit frontal image in Vivian 423r has another noteworthy aspect, the presence of the Bible. Upon presentation of the book, the ruler holds it, in all likelihood putting it on the left lap and touching it on the top with the left hand, in compliance with the exemplary image. The presence of the book, as entailed in Vivian 423r, poses then a compositional challenge of how to deal with the configuration of this new item and the extant scepter. As may be conjectured from the Pii drawing, their immediately resultant configuration would have been the positioning of the scepter on the top of the book, which is placed on the left knee. The extremely long scepter in Vivian 423r, however, does not fit the new configuration well: it would extend far above the ruler’s head, an awkward sight to look at and clumsy posture to hold steady indeed. Accordingly, the scepter would have been shortened, and probably reshaped so as to be like that depicted in the Psalter of Charles the Bald (Paris, Bibliothèque Nationale, Ms. lat. 1152), fol. 3v. Also of interest in this connection is the image preceding Vivian 423r, namely, Vivian 415v, the frontispiece to the Apocalypse: in it, St. John receives a measuring rod and a book from an angel. Although the rod here depicted is functionally different from the scroll/scepter, its relative length is comparable to that of the new scroll/scepter. Therefore, it may scarcely be ruled out that this neighboring picture served partly as an analogical basis for recomposition on the strength of a similar combination of a stick and a book. It may be interesting to observe further that in the illustration on Vivian 415v, with the rod having already been at John’s hand, the book is now being

48 

 2 God Enthroned: Pii/P2, with a brief discussion on P3.2b, P16.1, and P17.1

presented to him, the temporal order that is matched with the scenes represented in Vivian 423r. Turning to the positioning of the left hand, one would be led to presume, by assumption of minimal alteration, that, as an immediate consequence of the recomposition reconstructed in the preceding paragraph, the ruler held the book with his left hand at the top as well as the scepter—considerably shortened, though—at the bottom, the configuration that is closest to the one attested in Boethius 158v (Pl.  73), in which a scroll or a symbol of authority more generally conceived (Temple 1976: 59), rather than a scepter, is unambiguously represented. The exclusion of a possible identification of this object as a scepter in Boethius may be justified by the fact that it is definitely too short and too thick to be identified as such. In other words, the shortening goes much too far, exceeding the level reached in Pii and P2. Bringing this Boethius picture into comparison with Pii sheds a revealing light on the issue being treated. Of vital interest here is the correlation between the relative length of the rod-like item and the location of the hand. Where the rod is extremely short, the hand stays at the top of the book and at the bottom of the rod, as in Boethius. By contrast, where the stick remains long enough, the hand is shifted upward from the bottom to the middle section of the rod. Thus, either way, the placement of the hand at the interface between the book and the long rod does not materialize, exactly the configuration that would have resulted with a minimal alteration of the 423r image induced by the exemplar of the enthroned God. While this hypothetical configuration is not attested in the extant corpus, its actual occurrence or virtual reality may be inferred with reasonable confidence. That is, the former presence of the hand at the bottom of the current scroll/scepter, although not explicitly represented, may be plausibly deduced from the representation of the drapery springing out from the same place in the Pii drawing, in light of analogous depictions such as Grandval 352v (Pl. 25), Vivian 329v (Pl. 30), San Paolo 259v (Pl. 36), Dufay Gospels 179v, Le Mans Gospels 18r, and Lothair Gospels 2v, in which the hand is manifestly depicted and the comparable drapery hangs down from the cuff. At an earlier stage, then, the Lord’s left hand would have been placed on top of the book holding the bottom of the scroll/scepter. Be that as it may, one may hypothesize that the juxtaposition of the book below and the scepter above with the hand grasping at their juncture would have been found inconvenient to maintain steadily, probably because of its ill-balanced, precarious configuration. To remedy this physical inconvenience, two options became available. First, the scepter is shortened to such an extent that the two juxtaposed items can be held together fast. The excessive shortening, however, disfigures its original identity, thereby prompting a reidentification with something else, a scroll or more generally a symbol of authority, as in Boethius. Alternatively, having undergone shortening all the same, the scepter is left long enough to be still identifiable as a scepter, if also as something else. This option needs further modification in another respect: in order to support the still unstable configuration, the hand must be relocated to a higher position, the position that

2.2 A diachronic–comparative perspective 

 49

is manifested in Pii.35 Even this further remedial treatment does not completely remove the physical instability and aesthetic clumsiness of the juxtaposition involved. This would then have given rise to the conflation of the two distinct objects into a single rodlike item (whatever this may be conceptualized to be), concomitant with an increasing obscuration and an eventual loss of the identity of the book. This, I would claim, is the stage that is embodied in the current drawing of Pii. The peculiar rectangular portion immediately below the scroll/scepter as pointed out in the preceding section may be regarded as a trace of the original book, now obscured beyond categorical recognition as such. This redesigning may have been facilitated by P2, the paired companion image, in which the presence of a book would not have figured as prominently as in Pii, due to a greater degree of similarity to Vivian 423r in this regard, in which a book is not held by the enthroned king. Thus, the formal improvement of the original clumsy configuration of the scepter and the book has significant consequences in the conceptual domain by inducing a recategorization of the two objects. By virtue of its relatively less conspicuous size and shape, the book is less resistant to a conceptual obscuration that will ultimately result in a complete erasure from the scene. Accordingly, the final length of the original scepter on the one hand, and the resilience of the book to the bleaching of its cognitive basis on the other eventually determine a new categorization of the configuration at issue in a subsequent history of transmission and recomposition. In this respect, Junius 11 Pii and Boethius stand at opposite ends of the spectrum of evolution. In Pii, the old scepter seems to retain its initial identity at least in part, whereas the book is all but removed from depiction. By contrast, in Boethius, it is the identity of the scepter that has been disrupted due to its extensive deformation, thereby saving the book’s identity from being effaced. Leaving behind the depiction of the ruler himself and turning to his immediate surroundings in the model picture, one is confronted with the throne and the two court retainers, a configuration which may be conceptualized as a single unit (cf. Köhler 1933: 221), partly prompted by the grasping hands that link the otherwise three disparate beings. In fact, the suggested perceptual integration would likely have provided a cognitive basis for the overall symmetrical recomposition of the three discrete entities in Vivian 423r as the two attending seraphim in Pii. More specifically, the backrest and the two subjects are conflated into the two seraphim, whose entire wings constitute a substitute for a mandorla, an enlarged celestial backrest. Likewise, the footrest is made celestial through its replacement with the pair of cherubim. It may be interesting to note that the upper contour of the backrest in red—the curtain—in Vivian 423r is somewhat similar to the configuration of wings in Pii, as are its decorative elements resembling eyes. Thus, the red curtain of the throne and the clothes of the attendants were trans-

35 P2 differs in this regard, because the presence of a book is not a mandatory feature here, as discussed above.

50 

 2 God Enthroned: Pii/P2, with a brief discussion on P3.2b, P16.1, and P17.1

formed into wings of the seraphim, with a concomitant reconfiguring of the curtain’s decorative figures as eyes of the wings; correspondingly, the faces of the two retainers are redefined as those of the seraphim, while limbs and other body parts are left unexpressed for the reason elucidated earlier. Thus, the backrest, along with the two subjects standing next to the king on either side, is converted into the two seraphim standing just behind the Creator, while the footrest is reconstituted as two cherubim supporting the throne. The throne, conceptualized as an integral part of the ruler, absorbed his attending seraphim and cherubim into the unified whole of the Deity. The dedication portrait of the king is in this way made divine and converted to the representation of God enthroned. Above the throne hang curtains in Vivian 423r. The angled arch in Pii is formed in an analogous position above the upper wings of the seraphim. Curiously enough, the overall shape of the angled arch closely matches that of the array of the curtains. Given that the throne (the backrest) and the wings correspond to each other in compositional terms, as discussed above, it seems hard to resist the temptation to associate the arch and the curtains—the entities situated just higher than the compositional parallels in question—and to conjecture that the angled arch was due to a reinterpretation of the configuration of the curtains hanging from the arch in the model. As with other simplifying adaptations characteristic of Pii, the complex structure in the original—consisting of the arch and the curtains—would have been subject to reduction so that the arch in itself, to the exclusion of its decorative attachments, was retained, but transformed along the way through a partial amalgamation with those accessories. Moreover, the dominant yellow of the original arch and columns is reflected in the inner outline painted in a similar color, forming the angled arch and columns. In regard to the capitals, Vivian 423r contains two Corinthian order ones, which are depicted fairly precisely. In the Junius adaptation (Pii), they yield to considerable degeneration, to the point that a peculiar configuration obtains: what remotely resembles a frontal image of a pair of helices with a fleuron above them sits on a comparably corrupted form of an acanthus leaf represented in profile. Incidentally, the curtains hang down along the columns on both sides in the original. In the recomposed version of Pii, they are reinterpreted as part of the clouds, as are the outlines of the human figures standing in a semicircle in front of the king. Since these assembled people belong to the lowest layer in the original, the thematic concentration on the top level leads to its maximal marginalization resulting in the complete disembodiment of human individuals, leaving behind as their traces cloud-like silhouettes surrounding the Creator. It should be noted that clouds themselves are represented in the Vivian model, thereby providing a substantive basis of analogical extension for Pii. In the other recomposition (P2), given that the curtains on both sides hang in folds just behind the bodyguards in Vivian 423r, they presumably would have been reinterpreted by virtue of their proximity as hangings belonging to the men concerned. Based on such a reconfiguration, the right seraph and the leftmost angel came to be depicted with drapery in P2.

2.2 A diachronic–comparative perspective 

 51

Before bringing this section to an end, it may be appropriate to comment briefly on Raw’s (1953) view on the influence of the Utrecht Psalter on Pii and that of the Roda Bible (Biblia Sancti Petri Rodensis, vol. 3; Paris, Bibliothèque Nationale, Ms. lat. 6) on P2. Assuming that the representation of the winds in the form of human faces constitutes a prime defining feature of Pii, Raw (1953: 68–69) presents three pictures in Utrecht, namely, Utrecht 9r (Pl. 74), 59v (Pl. 79), and 76r (Pl. 80), as possible sources of the Junius picture in question. While the personification of the winds commonly underlies all four drawings, none of the three Utrecht images can be viewed as even remotely parallel to the Pii representation in iconographic details. The Junius winds are represented as heads just beneath the feet of the throne, provided with large wings on both sides. These conspicuous wings are found nowhere in the Utrecht pictures. Nor are the heads situated below the throne, as none of the three Utrecht pictures has a throne depicted. Furthermore, not a single instance among the three pictures embodies the clustering of other properties that seem to be as distinctively characteristic of Pii as the unique representation of the winds. At issue are “being seated on the throne” (Pii) rather than standing (9r) or sitting on a globe (59v, 76r), “a full-face view of God” (Pii) rather than a profile (9r, 59v), “outward-facing six-winged seraphim” (Pii) rather than inward-facing (9r, 59v, 76r) or double-winged angels (59v, 76r), and “without a mandorla” (Pii) rather than with a mandorla (9r). Invoking the three separate images as bases for an eclectic combination of their attributes for the composition of Pii would hardly enhance explanatory power, insofar as the actual cooccurrence pattern fails to derive from a feasible course of development. Most significantly, in none of the Utrecht images do even traces of architectural elements such as pillars and arches appear, which distinctly serve as frames of the Pii picture (Reinhard 2010). Thus, the postulation of the Utrecht images in question as principal sources of Pii is incapable of offering a cogent explanation for the elaborate architectural features of the Junius picture. In light of these outstanding disparities between Pii and the Utrecht pictures, it may be warranted to infer that the influence of the Utrecht Psalter on the production of the Pii image would have been marginal, if at all. After all, far from unique to Utrecht, the configuration of the Deity flanked by seraphim is obviously widespread, as is exemplified also in Vivian 30v and San Paolo 117r (Pl. 35). And the notion of God riding on the cherubim/winds would have been no less familiar, as noted in the previous section.36 As regards P2, three pictures have been adduced as its possible model with varying plausibility: (i) the vision of Isaiah, the Roda Bible, fol. 2v (frontispiece to the Book of Isaiah); Raw 1953: 69–70; see also Broderick 1978: 76–77; (ii) the blessing of the seventh 36 Of related interest is Utrecht 47r (Psalm 79). Indeed, this image of God flanked by seraphim displays a striking similarity to Pii, as correctly observed by Reinhard (2010): God is seated with his left leg raised higher than the right. In other respects, however, this illustration differs no less markedly from Pii than the three Utrecht pictures treated above: enclosed in a mandorla, God faces left in three-quarter view; the throne is not explicitly depicted, nor are the cherubim/winds represented; the seraphim look inward to God; and no architectural frame is provided.

52 

 2 God Enthroned: Pii/P2, with a brief discussion on P3.2b, P16.1, and P17.1

day, the San Marco mosaics (Pl. 39); Broderick 1978: 86–89; (iii) the man-child rescued by God, in an array of illustrated manuscripts of the Commentary on the Apocalypse by Beatus of Liébana, a Spanish monk who died in 798 (Ohlgren 1975: 65n12; Williams 1994–2003.1: 13–18); the relevant manuscripts include the Morgan Beatus (New York, Pierpont Morgan Library, M. 644, fol. 153r; Williams 1994–2003.2: 21–33) and the Valladolid Beatus (Valladolid, Biblioteca de la Universidad, MS 433, fol. 131r; Williams 1994– 2003.2: 38–42) as the earliest extant copies; Ohlgren 1975: 41–44; see also Broderick (1978: 77–78); for fuller discussion on a more solid basis drawing on Williams’s comprehensive corpus, see section 12.5.2 below. Common to all these works, God is enthroned, flanked by angelic beings; and to the lower left of the Creator is depicted the main character of the scene: (i) Isaiah, whose vision is being narrated (Roda; Isa. 6:1–3); (ii) the angel representing the seventh day being blessed by God (San Marco; Gen. 2:2–3); (iii) and the man-child who has been brought before God for protection from the dragon (Beatus; Rev. 12:1–5). The above order in which the three possible sources are presented corresponds to the scale of increasing convincingness that may argue for a borrowing through contact. The least feasible is the Roda Bible, in light of the ubiquitous occurrence of the representation of God in Majesty, which is too general an image to warrant postulating a specific inspiration from this work at the expense of numerous other candidates. In the absence of a plausible path of derivation proposed in lieu of the vague speculation (Raw 1953: 70), the asymmetrical arrangement of seraphim in P2 would be hardly compatible with the perfect symmetry in the alleged model. No less divergent is God’s three-quarter view in P2 as opposed to the frontal image maintained in the Roda Bible, even though an additional figure (Lucifer or Isaiah) is depicted comparably in the lower left. Also at variance is the presence (Junius) versus absence (Roda) of another angel behind the left seraph and the main figure. Brought together, then, P2 and the Roda Bible are distinguished sharply by the radical difference in the extent to which symmetry is realized in overall composition. Given that a symmetricalization of an earlier, asymmetrical composition seems more likely to occur than a contrary process of disharmonization, Raw’s hypothesis would require specific justification in terms of derivational mechanism and trajectory. Furthermore, the alleged inspiration seems to lack a conceptual basis: it is far from evident how the vision of Isaiah associates with the Fall of the Rebel Angels. Last, but not least, given that there is no other case in which the Roda Bible is invoked as a model for Junius 11 (the first twenty-two pictures), the appeal to such a low-profile source would appear suspicious and infeasible. Much the same criticism applies to the Blessing of the Seventh Day image at San Marco. The line of derivation from more symmetrical to less symmetrical representation, as is entailed, would necessitate particular verification. The association of the Sabbath with the Fall of the Angels remains to be substantiated to make the purported image transferal between the two distinct conceptual domains a likely event. Compared with the Roda Bible, it might seem more plausible to posit a genealogical correspondence between Junius and San Marco here, given that the two are demonstrably derived

2.2 A diachronic–comparative perspective 

 53

from common sources on independent grounds (section 9.2). Lacking empirical specificity in this instance, however, it is scarcely warranted to accept the claim without demonstration. In contrast to the previous two cases, Beatus seems far more likely to have played a formative role, not primary, to be sure, but at least auxiliary, as argued below. Most significant of all, there are a few specific similarities in evidence between Beatus and Junius unlikely to have developed independently and accidentally. These notable resemblances concern a layout on the page. As may be recalled, one peculiarity of P2 is that the image is presented at the upper right periphery of the picture zone. Interestingly enough, the scene of the rescued man-child is found at the corresponding place, the upper right corner of the page. Moreover, the man-child, exactly like Lucifer, appears at the center of the scene, with the figure of the enthroned God relegated to the right margin. Still further, just as the fighting scene appears to the right of P2 on the facing register of the following page (P3.2b; section 3.1), a similar representation can be seen to the left of the man-child scene. Down below at the bottom, the rebel angels are represented falling, fallen, and bound in hell, in Beatus and Junius alike (provided that P2 and P3 are taken to be continuous). Although these resemblances may be attributed in part to the conceptual commonality of the Fall of the Angels in the Old English/Old Saxon Genesis on the one hand and in the Book of Revelation on the other, the common thematic basis hardly can explain why the distinct scenes involving Lucifer and God on the one hand, and the man-child and God on the other, must be depicted at the upper right corner of the available space in both cases, with Lucifer/the child at the center and God at the right. Confronted with such conspicuous agreements in spatial arrangement, one would have no choice but to acknowledge their affinity in this respect. Indeed, the close resemblance between Beatus and Junius in the layout of the pictures of the Fall of the Angels is beyond reasonable doubt, but it is limited to the overall placement of figures and scenes. In other words, the pictorial agreements between Junius and Beatus hardly extend to specificities in individual figures and configurations, in sharp contrast to the relationship between P2 and Vivian 423r. As argued at length above, numerous fine details not only of P2 but also of Pii become subject to explanatory accounts by identifying the two pictures as covariants derived from Vivian 423r as an underlying model. In view of the quantity and quality of specific correspondences involved, Beatus’s inspiration as epitomized in the Morgan Beatus 153r for P2 must be regarded as secondary to that of Vivian 423r.37

37 For further discussion on the similarity between P2 and the man-child scene in the Beatus Commentaries and its implications for the Touronian foundations of Junius, see section 12.5.2 below.

54 

 2 God Enthroned: Pii/P2, with a brief discussion on P3.2b, P16.1, and P17.1

2.3 Residual anomalies Building on the assumption that Vivian 423r and the exemplary representation of God enthroned in the School of Tours provided sources in their own different ways for the drawings of Pii and P2, the preceding two sections—from complementary perspectives, synchronic–structural and diachronic–comparative–genealogical—identified the complex of structural and cognitive factors leading to new organizations and compositions in Junius 11, and inferred with reasonable plausibility the ways in which the source materials were adapted and reconstituted effectively to suit the artistic motivations and purposes. Although the two drawings are designed and executed largely with admirable consistency in accordance with the given sources and the artist’s concerns and interests, there remain a number of flaws and inadequacies in both pictures, some of which have already been mentioned in passing. This section brings these failures altogether and explores their implications for a better appreciation of Pii and P2 in structural and art historical dimensions. As far as Pii is concerned, only one inadequate treatment seems to be detectable. The left eye of the seraph standing to the right of the Creator is depicted on the wing, outside the contour of the face. Without the left eye, this seraph would be perceived as properly represented in profile. As it is, however, the three-quarter face that is formed by the left eye at issue conflicts with the side view that is shaped by the outline of the head. In light of the close parallelism inversely obtained between the seraphim’s faces in Pii and P2, pointed out in section 2.1 above, this apparent flaw may be viewed as an analogical imitation of the left seraph in P2, for which the eye on the wing depicted closest to the seraph’s own (right) eye could be construed as its left. If so, the apparent error would have been a deliberate operation. In sharp contrast, P2 suffers no fewer than five errors in composition. First, the lower outer wing of the seraph standing to God’s left contains a blank section that is located just behind God’s right fingers and the seraph’s own thumb. This in fact must be covered with feathers, as with the portions immediately above and below it (Fig. 2). Rather than a simple oversight, this inadequacy might have been deliberately made to avoid the otherwise resulting overlapping of two distinct entities, as will be argued below. Second, the upper left wing of the right seraph must be decorated with eyes, rather than waves, in light of the pattern displayed by the left seraph, and also according to the complementary distribution of eyes and waves between Pii and P2, as pointed out in section 2.1 (Fig. 2). Presumably, the partial sharing of feathers (waves) by the opposite wing, decorated with a wavy line and rows of eyes, would have misled the artist into executing the same decoration of wavy lines throughout in the left wing (for an elaborated account, see below). Third, the lower right wing of the same seraph must be feathered, rather than being left blank, in harmony with the facing wing and also with the corresponding decoration of the opposite seraph—again, as dictated by the complementary distribution of eyes and feathers at issue (Fig. 2). Fourth, the portion between the Lord’s gripping fingers and the tapering drapery must be left plain, rather

2.3 Residual anomalies 

 55

than being feathered, because this section should be part of the Lord’s left hand hanging (Fig. 2). Again, this oversight might be regarded positively as another instance of avoiding overlap. Putting the third and fourth points together, the area to the left of the scroll/ scepter and the one below it rightward must be reversed in their filling, lined on the one hand and plain on the other. An alternative reading of the seraph holding the scroll/ scepter, however, would justify the depiction of feathers here. Yet, while firmly based on Vivian 423r and thus plausible enough in diachronic–comparative terms (section 2.2), this interpretation, seen from a strictly synchronic perspective, may seem rather contrived, if not utterly improbable, as noted in section 2.1 above. Fifth, the feathers of the lower left wing of the right seraph must be longer; accordingly, the blank space—the triangular zone of which the feathers below the (Lord’s) fingers and the scroll/scepter form two upper sides—must be filled with feathers (Fig. 2). Furthermore, the right seraph not only has fewer wings (four in all, as opposed to six in normal cases), but also is too short in height, as it has its lower part of the body below the waist unrepresented at all, despite there being space available for illustration. Also illuminating is a comparison with the two angels depicted on the far left (Pl. 2). To be sure, they have no feet represented, but they are depicted down to the knees. The abbreviating treatment of the right seraph on P2 is indeed quite unique.

Figure 2. Junius P2 (detail), reconstructed (left) and actual (right). Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 2, God Enthroned, Facing Lucifer (detail, adapted). Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).

Thus, while there is apparently one error in Pii, there are as many as five in P2. Such a numerical contrast proves to be statistically significant. Comparing the right seraphim in Pii and P2—the major recipients of errors in either drawing, as will be demonstrated shortly—on the basis of whether their wings are each depicted correctly or incorrectly,

56 

 2 God Enthroned: Pii/P2, with a brief discussion on P3.2b, P16.1, and P17.1

one gets the following contingency table:38 for Pii, five correct wings versus one incorrect wing; for P2, one correct wing versus five incorrect wings. Running a Fisher’s exact test, one gets a p-value of 0.080. It may therefore be concluded that P2 is more likely to display errors than Pii with a statistical significance at a level of 0.10. Furthermore, with a single exception, the errors in P2 are concentrated in the right seraph. This right–left asymmetry in error occurrence is also of statistical significance. Taking into account the maximal constituents of the seraph to be illustrated (head, two hands, two legs, and six wings), one may determine the relative likelihood of correct representation, based on the following contingency table: for the right seraph, nine incorrect representations (including missing body parts) versus one correct representation; for the left seraph, one incorrect representation versus nine correct representations. Performing a Fisher test gives a p-value of 0.001, which justifies a conclusion that the right seraph is more susceptible to errors in P2. By contrast, a similar statistical analysis for Pii results in no significant difference between the right and left seraphim in their likelihood of errors: for the right seraph, five incorrect representations versus five correct representations; for the left seraph, four incorrect representations versus six correct representations. The p-value obtained is 1, meaning that there is no difference at all between the two seraphim in Pii in their likelihood of being affected by errors. In addition to the quantitative differences disclosed above, a further outstanding distinction between the Pii right seraph and the P2 counterpart must be mentioned: the only error detected in Pii is a matter of a single stroke or two, notably an outline curve of the forehead. Thus, Pii is not only less likely to receive errors, but also the exceptional error itself is minimal compared with those occurring in P2, which concern a complex of operations, specifically a filling of a given space. The absence of the limbs as well as the two outer lower wings in the right seraph of P2 may to some extent be ascribed to the lack of space for a full illustration of the figure, as argued in section 2.1. The unavailability of sufficient space is in turn a consequence of the centralization of one of the attendants in Vivian 423r—now Lucifer—and the concomitant decentralization of the Creator. Nonetheless, a full explanation has yet to be explored, as there still remains space sufficient for a partial depiction of many of the missing elements. More seriously, all of the six remaining deficiencies, including the excessive trimming of the lower body, are not readily susceptible to a similar explanation: despite the space available in the current layout, a proper illustration is not executed. A complementary account accordingly would be needed to be formulated. In fact, there is one available, the one that is derived from the focus on Lucifer and his relation to the Lord, the major concern of P2. In light of the drawing’s primary interest in the God–Lucifer line of sight, it would be a natural consequence that the farther a given figure is located from this axis, the less important it becomes, and hence

38 For the sake of statistical analysis, the missing two wings in the P2 seraph are counted as incorrect representations.

2.3 Residual anomalies 

 57

the less salient it is represented. Since the left seraph is close to the axis, it receives a full representation. By marked contrast, the other seraph, which stands at the opposite end of the group and therefore farthest away from the axis, is subject to a maximal defocusing and obscuration in representation. Since this obscuration is a perceptual/ cognitive matter not entirely reducible to the physical dimension of space, the resort to this independent factor may provide an explanatory account for the extensive trimming that the right seraph suffers, including the total absence of the lower part of the body, the outer wings, as well as limbs, regardless of and complementary to limited space availability. Of related interest, the replacement of eyes with waves in the upper left wing might also be characterized as an obscuration subsequent to defocusing. Eyes on the wings are distinctive features of seraphim; accordingly, the maximally diminished prominence that falls on the right seraph may imply a simplification of this markedly decorative feature to waves, the unmarked decorative element of a wing. By transitivity, yet a further weakening ultimately may lead to a complete removal of this less distinctive property, as exemplified by the plain lower right wing. Such fading effects on defocused figures make themselves felt not only on the surface dimension, but also on that of depth. To be more specific, background objects tend to be obscured in favor of the foreground ones even to the point where they may be left totally undepicted. In this light, the blank area for the left wing behind the Lord’s right hand, and the lack of the lower part of the left wing of the right seraph lend themselves to a principled explanation. Of paramount interest is the former case: the middle section of the outer wing intentionally would have been left plain to give extra prominence to the Lord’s right hand, positioned over that portion of the wing and pointing toward Lucifer. As it turns out, the avoidance of overlapping that has been referred to above may, in the final analysis, be reduced to local effects of pictorial focusing/defocusing. One is then left with the feathers of the upper portion of the hand drapery as a still unexplained property. In the presence of such a seemingly intractable flaw, one might be tempted to conclude that it was after all a mistake, rather than a calculated deviation. On closer consideration, however, this error unlikely would have been a random one due to simple carelessness; rather it probably would have been a deliberate alteration based on the artist’s misconceived idea that the drapery belongs to the seraph, as is the case with the leftmost angel (Pl. 2). It may be instructive to compare also with Pii, in which the corresponding drapery is obviously dissociated from God’s hand (see section 2.1 above).39

39 Of related interest is an Ottonian MS, Bamberg Msc. Bibl. 76, fol. 10v (frontispiece to Isaiah; Hieronymus, Sophronius Eusebius: Isaias glossatus), in which the seraphim, six in all, display drapery under the wings. Cf. Raw (1997: 135). It might be speculated that in the course of transmission the drapery originally associated with the Creator’s hand was reconfigured as part of a seraph’s wing, a reinterpretation based on a representation in the late tenth or the early eleventh century.

3 The Fall of the Rebel Angels: P3 3.1 A synchronic–structural perspective The overall composition of P3 (Pl. 3) is organized in complex, multi-tiered, nested ways, as follows. It comprises three major registers (P3.1, P3.2, and P3.3), each of which in turn is divided into two layers (P3.1a, P3.1b, P3.2a, P3.2b, P3.3a, and P3.3b). In P3.1, the division is incomplete, with only the right side being double-layered. In P3.2, the subregisters are most clearly demarcated. In P3.3, the two layers, the air and the gate to hell, are not formally divided, yet they are demonstrably distinguished in temporal and functional terms implied by the pictorial representation involved. The division between P3.1a and P3.1b and that between P3.2a and P3.2b are of lesser importance compared with the major boundaries between P3.1, P3.2, and P3.3. The varying degrees of demarcation at work are indicated by the fact that the feet and wings of angels, and the javelins held by God trespass on the weaker boundaries, but not on their primary, stronger counterparts. Among the three main components, the first two major panels (P3.1 and P3.2) are grouped together, as they are representations in heaven, whereas the last one (P3.3) is a depiction out of heaven. This bisection is marked out by the thickest border on this page, located between P3.2b and P3.3a and supplied with an inscription written in red, “hēr se” (here it [is]/here [is] the [thing]; compare Mitmann/Kim 2015: 6).40 Thematically, however, the first three minor panels—P3.1a, P3.1b, and P3.2a—describe the rise of Lucifer, while the last three represent his fall, or in more precise terms his failure in heaven and fall into hell. The thematic division thus bisects the six subregisters into halves, the bipartition that is specifically implemented by the molded border—the only one of this kind—standing between P3.2a and P3.2b. According to the traditional conceptualization, P3 is subject to either a tripartite (Gollancz 1927: xxxix; Raw 1976: 143–144) or quadripartite (Broderick 1978: 93; Ohlgren 1992: 89) division. Common to these two alternatives is an exclusive reference to the overt boundaries—the ones between P3.1 and P3.2a, between P3.2a and P3.2b, and between P3.2b and P3.3)—in defiance of the formally implicit, or less than full,

40 In the wake of the caption “Hēr se hǣlend gescē[op] helle heom tō wīte” (Here God created hell as a punishment for them), provided on the right margin between P3.2a and P3.2b, this fragmentary inscription, her se, may be interpreted as “Here (is) the (result)/Here (is) the (hell of God’s creation).” Thus, the truncated expression emphatically signals that an entirely different world from heaven begins here and now in the presence of the viewer/reader. It should be noted in this connection that, while heaven and hell are unbridgeable, heaven and earth, and hell and earth, are passable from one to the other, as represented in P9, P20, and P36. To generalize following Mittman and Kim’s (2015: 7, 11, 18–20) illuminating account, the inscription fulfills the pragmatic–deictic function of presenting before the viewer’s/reader’s very eyes the critical moment of the transformation from Lucifer in heaven into Satan in hell. https://doi.org/10.1515/9783110788068-003

60 

 3 The Fall of the Rebel Angels: P3

delimitation that I have characterized as significant above. Either partition, then, partly overlaps and intersects with my threefold double-tiered organization. While commonly recognizing P3.1 as a single domain (which I conceive as comprising two subdomains on the right side), the tripartition divides between P3.2a and the remainder (P3.2b/P3.3), whereas the quadripartition gives an autonomous status to each of the remaining three panels that are demarcated in an explicit manner (P3.2a, P3.2b, and P3.3). Gollancz’s and Raw’s grouping together of P3.2b and P3.3 despite the categorical boundary located in between obviously rests on his (and her) committed assumption that the two scenes are inseparably united, as stated explicitly in his description: “The upper part represents the wrathful Deity, attended by angels, holding three javelins with which He strikes downward [italics supplied]” (Gollancz 1927: xxxix). Inasmuch as God is conceived as aiming his javelins at the falling angels, notably Lucifer, who appears to be depicted directly beneath the tips of the javelins, it may be reasonable to treat P3.2b in unison with the panel below. On closer reexamination (see further below), this interpretation proves to be far from self-evident and is superseded by an alternative account that does not require the compositional unification that Gollancz took for granted. Thus, Gollancz’s tripartition must be rejected because it falsely postulates a visual unity and simultaneity where none exits.41 In contrast, the quadripartite division should not be wholly accepted for the contrary reason of failing to see a conceptual continuity/sequentiality lying behind P3.2a and P3.2b, as will be discussed in due course. In fairness, however, insofar as this division is taken to be simply a descriptive first approximation based on the use of manifest boundaries, there will be no serious harm in referring to the overt four-panel construction of P3. Proceeding now to individual examinations of the pictures in the major registers, the boundary between P3.1a and P3.1b on the right side seems to suggest that the two panels involved be read independent of each other, while the large figure of Lucifer, which extends to the two subregisters, may well be interpreted as being connected with both. On the one hand, glancing backwards and beckoning, Lucifer invites the four angels in the upper panel to the new palace, resulting in the entry into the hall as depicted on the left side of Lucifer. On the other, Lucifer receives (or is about to receive) crowns from the lower procession of angels, who all look up to him. That he already wears one constitutes a conceptual link, implying a causal relation, to the crown-bearing angels, as does the part of the throne—a footrest or a pedestal—represented under Lucifer’s feet. Furthermore, his holding the scepter implies that he is formally treated as king, a status that must be consecrated by a coronation and enthronement to be held inside the new palace. Accordingly, part of Lucifer’s representation anticipates what is to occur immediately afterwards in the palace. More specifically, the empty throne on

41 Needless to say, the temporal distinction as successive scenes does not gainsay that P3.2b and P3.3 unite at a more abstract, thematic level, namely, the start and end of Lucifer’s fall, as pointed out above.

3.1 A synchronic–structural perspective 

 61

the left and the one immediately behind Lucifer may well be identified as one and the same.42 Thus, the footrest under Lucifer may be interpreted as corresponding to the one under the throne in the palace: it would have been intended as an abbreviated representation of the throne, with an exclusive focus on the standing Lucifer at the expense of surrounding details. In this way, there is no implication of a direct relation between the upper and lower rows of angels: they are incorporated into a larger configuration as its constituents through intermediacy of the central figure, Lucifer, with whom they are meaningfully associated in their own ways.43 Naturally, these two scenes should be interpreted in temporally differentiated terms, rather than constituting two separate events occurring at the same time: the lower scene occurs in succession to the upper one after Lucifer and his attendants enter the palace, as depicted on the left side. Following the coronation, he receives (or distributes) palms, as represented in P3.2a (see below). On the whole, then, P3.1 depicts an ongoing procession to the new palace and a subsequent occupation of the throne. Accordingly, ambivalent as it is, the central figure belongs to the two separate scenes that are temporally and spatially distinct: intervened by the entrance scene on the left side of P3.1, they take place outside and inside the palace, respectively. While Lucifer’s posture, notably his head turned behind and his betokening right hand, represent the approach to the palace, the footrest under Lucifer implies his occupation of the throne, an enthronement that is to take place in the new palace. That Lucifer and his vassals actually have entered the palace—rather than envisioned in Lucifer’s mind—is confirmed by his fall, together with pieces of his throne, in P3.3a (see below). The group of angels apparently behind the building represented on the left side confirms the interpretation that the band of angels on the upper right side is urged by Lucifer to enter the building, rather than simply being directed to observe it from outside. One might go further to assume that the angels on the right side are actually proceeding past Lucifer at the center to arrive at the palace. The two groups accordingly may be identified with each other; the same angels are depicted twice at the end as well as at the start of their march to the palace. More minutely on the left group at issue, the four angels beyond the roof can be viewed as standing beside the throne inside (rather than outside) the palace, as do the king’s vassals and bodyguards—the four attendants—in Vivian 423r (section 2.2). More specifically, the scene on the left side in P3.1 represents the situation in which the angels are ready to welcome Lucifer to occupy the throne by partially replicating P2, in which three angels face right and one faces left, exactly as with the group of angels in P3.1 42 Incidentally, the location of the throne on the second floor constitutes an iconic representation of the verse phrase hēahran on heofonum ‘higher in heaven’ (Genesis B 274a). 43 Such a conflated representation of two distinct scenes in the single figure of Lucifer is strikingly analogous to that of Satan’s messenger in P28 (section 8.1) and, ultimately, reminiscent of the technique executed in Vivian 423r (section 2.2).

62 

 3 The Fall of the Rebel Angels: P3

under consideration.44 This interpretation will be strengthened from a comparative perspective in section 3.2. Correspondingly, the fifth angel counted from the extreme left in P3.1, just behind the central figure of Lucifer, may be identified as Lucifer, who is about to occupy the throne, as argued below. The crowned Lucifer with a scepter, along with the procession of angels offering crowns, presumably refers to the coronation that takes place in front of the new throne after Lucifer and his followers have entered the new palace, as depicted on the left. In this way, Lucifer’s figure, represented singly at the center of the register, may be viewed as a conflation of Lucifer involved in the three different scenes that unfold sequentially: the directing Lucifer (part of P3.1a on the right), the escorting Lucifer (part of P3.1a on the left), and the Lucifer being crowned and to be enthroned (part of P3.1b).45 By the same token, the two rows of angels standing in the double layers on the right side might possibly be regarded as one and the same: the angels outside and inside the palace, respectively. At this point, one is confronted with the apparent asymmetry between the groups on the right and left: while there are four angels on the right, there are as many as five depicted on the left. Although a perfect match between the two sides is far from a logical necessity of the reading proposed above—thus, this does not count as crucial counterevidence against the claim being adduced—one may speculate on an alternative interpretation to make better sense of the disparity at issue. The rightmost angel in the left group—the fifth counted from left—can be viewed as Lucifer himself, as suggested above. After the four angels move to the palace past Lucifer by his summoning, Lucifer follows them from behind. Accordingly, in the next scene represented on the left side, he stands on the rightmost, preceded by the four angels. The group of five angels on the left side is thus the same as the group of five on the right, headed by Lucifer in the center. The prominent size of the angel in question may reflect his relative importance in the left group, in parallel to Lucifer at the center relative to the right groups of angels, above and below the bordering floor. As may be recalled, the crown and the scepter fully fit the ensuing ceremony to be held inside the palace. Therefore, at the time of entry, Lucifer would not yet have worn either of the sovereign insignia. The double occurrence of Lucifer in the second (P3.2) and third (P3.3) registers adds plausibility to the interpretation proposed here whereby Lucifer is represented (explicitly) twice in the first register as well. In summary, P3.1 represents three separate scenes, differentiated as they are in space and time. The first two take place sequentially one after the other outside the palace, starting at the right and ending at the left in the upper half of the register. They 44 Consider also a parallelism with P2, which, immediately preceding P3.1, associates with it conceptually and compositionally (section 2.1). Immediately following the scene of God enthroned attended by his closest company in P2, P3.1 in general and P3.1 (left) in particular represent a polar opposite of God’s court, headed by Lucifer. 45 This suggestion of the single figure of Lucifer playing the triple roles will be modified below.

3.1 A synchronic–structural perspective 

 63

are thus the initial and final phases of a continuous action of entrance into the palace. By contrast, the last scene is set inside the building, as Lucifer is shown standing at the throne, in the process of being crowned and enthroned. While one might criticize my indulging in an outrageously complicated reading of the seemingly straightforward picture of Lucifer attended by a large host of angels, there turn out to be some empirical grounds for substantiating this interpretation when viewed from a comparative perspective, as will be demonstrated in due course (section 3.2). At this juncture, I confine myself to the synchronic–structural dimension and raise a few issues with the “minimal” reading as exemplified in Raw (1976: 144), Broderick (1978: 91), and Ohlgren (1992: 88). In their view, Lucifer is surrounded by three groups of angels totaling thirteen, implying that the scene is monolithic and undifferentiated in spatiotemporal terms. This atemporal and holistic conceptualization entails that the whole register (P3.1) be organized throughout the canvas uniformly, without significant local distinctions in time and space. This implication is not borne out on structural grounds. In fact, the first register of P3 (i.e., P3.1) is far from uniform and homogeneous regarding its overall pictorial organization. In specific terms, this register (P3.1) is composed of five major components, whereby the central member, Lucifer, is encircled by the four peripheral ones—the three separate groups of angels and the building. These constituents are organized into the whole representation with varying degrees of cohesion by means of explicit pictorial devices—gaze and/or gesture—and also by depiction of associated items. The axis of Lucifer at the center, the upper angels to the right, and the palace to the left constitute a dense network, whereby, beckoning towards both ends, the central figure looks back to the angels, who face squarely toward the center. By contrast, the axis diagonal to this major configuration, which stretches from the upper left to lower right groups, is less than fully articulated, with a lesser use of explicit reference and a greater dependence on implication for a coherent interpretation. Least fully incorporated in this minor axis is the upper left periphery, the group of angels seemingly behind the palace, who are oriented in diverse directions contrary to the other two bands aligned in parallel. The isolation of this component is most tellingly testified by the absence of meaningful markers connecting the two immediately proximate figures, Lucifer and the relatively large angel to the left—whom I have identified as Lucifer in the above discussion. Despite the direct contiguity involved, no sign of contact seems in evidence between these two figures, who turn their back on each other, as if they were mutually invisible. In the absence of any explicit connection with the central figure, and to say nothing of their own extreme peripherality, the five angels in question will remain beyond understanding insofar as they are conceptualized as synchronous with the other bands of angelic beings. Less fragmented, but still not to be ignored is a lack of direct coordination between the two rows of angels on the right side—only the lower angels carry offerings. This structural discord also casts doubt on the synchronicity of the two-tiered procession. Finally, the temporal sequencing of the constituents within the other two major regis-

64 

 3 The Fall of the Rebel Angels: P3

ters, P3.2 and P3.3, lends plausibility to an analogous sequential interpretation internal to the first register, P3.1, as well. As regards the picture in P3.2a, its proper interpretation hinges on the identification of the central angel, who receives palms from (or distributes them to) his fellow angels to the right and the left: with whom is he identified, a rebel to God or his supporter, or more specifically, Lucifer or Michael? According to the dominant view (e.g., Gollancz 1927: xxxix; Ohlgren 1992: 89; Mittman/Kim 2015: 3), the prominent angel is identified as Lucifer, who receives palms as tokens of victory from his followers. P3.2a thus represents Lucifer at his zenith, comparable to God in heavenly status. In reaction, taking issue with this standard interpretation, Rosenthal (1974: 190–191) —endorsed by Broderick (1978: 102–103) and reiterated by Karkov (2001: 50n14; although without acknowledgement to Rosenthal)—suggests that the angel at issue ought to be grasping palms as symbols of victory and triumph on behalf of God, who has just triumphed over Lucifer, as depicted in the register immediately below. Thus, P3.2a, representing a celebration of God’s victory, constitutes a scene consequent on P3.2b. Inasmuch as the middle-standing angel serves as an acting head of God’s host, he must be regarded as his closest attendant, that is, by implication, Michael. Following Rosenthal’s argument, since Lucifer has never triumphed over God, the attribution of palms to Lucifer that the standard identification of the central angel entails is infeasible on conceptual grounds. In the absence of textual reference, explicit or implicit, to Lucifer’s fight against God, to say nothing of his victory, Rosenthal (1974: 191), and Broderick (1978: 103) in his wake, go on to reason, to associate a symbol of triumph with Lucifer would be extremely unlikely in contextual terms. As it is, however, the interpretation as Lucifer seems to fit most naturally the text running from Genesis B 283b through 289a: ic mæg wesan god swā hē.  // bigstandað mē strange genēatas,  / þā ne willað mē æt þām strīðe geswīcan, // hæleþas heardmōde. / hīe habbað mē tō hēarran gecorene, // rōfe rincas. / mid swilcum mæg man rǣd geþencean, // fōn mid swilcum folcgesteallan. / frȳnd synd hīe mīne georne, // holde on hȳra hygesceaftum. / ic mæg hyra hēarra wesan, // rǣdan on þis rīce. (I can be a god as he is. Strong companions stand beside me, tough-minded warriors who will not fail me in the strife. They have chosen me as their master, the brave solders; with such supporters one can consider plans, undertake them with such as these. They are my zealous friends, loyal in their hearts. I can be their master, govern in this kingdom.) (Anlezark’s [2011: 23] translation)

In contestation, one might invoke a compositional parallelism of P3.2a/P3.2b to P9.1/P9.2 to strengthen the above nonorthodox claim. Specifically, it seems that P3.2a is to P3.2b as P9.1 is to P9.2. In both pairs, it may be generalized, a host of angels headed by Michael stand in the background, and God acts in the foreground. This alleged compositional parallelism might support the identification of the central figure in P3.2a as Michael, because the corresponding figure in P9.1 is designated Michael in the accompanying inscription (section 5.1). On closer inspection, however, there is a notable difference. While the two registers in P9 are simultaneous scenes on earth and in heaven, the cor-

3.1 A synchronic–structural perspective 

 65

responding two in P3.2 are not; rather, P3.2a is conceptualized as consequential to P3.2b, as noted above in regard to Rosenthal (1974). Moreover, the two scenes involved both take place in heaven. Therefore, the alleged isomorphism between P3 and P9 cannot appropriately be adduced as decisive evidence for identifying the central figure in P3.2a as Michael. The identification of the pivotal figure as God’s major proponent is further weakened by the disparity between the alleged temporal order of events and the arrangement of the panels in P3.2: the sequence P3.2a and P3.2b is contrary to the allegedly actual course of events, which should be P3.2b followed by P3.2a. Given that the spatial arrangement parallels the temporal sequencing in the other registers, the apparent reversal in order here may call into question the interpretation under consideration. More seriously and substantively, a clustering of compositional properties may be adduced that, once brought together, argue for identifying the central figure in P3.2a as God’s antagonist or Lucifer in particular, rather than God’s adherent or Michael in particular. First is the overall framing of the pictures involved. Specifically, the location of the capital in the right pillar suggests that the first (P3.2a) and second (P3.2b) subregisters of the second major register (P3.2) would have been designed as an integrated scene in distinction from the registers above (P3.1) and below (P3.3), as does the ground that is segmented or jointed in the middle section between P3.2a and P3.2b, in contrast to the continuous border lines that separate between the first and second major registers (P3.1a/ P3.1b vs. P3.2a), and between the second and third major registers (P3.2b vs. P3.3a). Second, in the top and bottom registers, Lucifer makes multiple appearances. In P3.1, Lucifer, who occupies the space that corresponds with both the upper and lower subregisters at the same time, is thus represented twice implicitly, outside and inside the palace, as argued above. Further, under the interpretation that the rightmost angel in the upper left group is identified as Lucifer, he is depicted twice in explicit form as well. Most transparently, Lucifer is represented twice in P3.3, as a falling Lucifer and as the bound Satan, respectively (see below). Given that the middle two panels (P3.2a and P3.2b) are subsumed under the second major register as its integrated components, as pointed out in the last paragraph, it seems plausible to assume that Lucifer may receive comparable treatment through his double appearance in the middle register as well, first in P3.2a and for the second time in P3.2b (as substantiated below). Altogether, then, it may be generalized that Lucifer (or Satan) is represented consistently in each of the six component panels of P3. This ubiquitous presence of Lucifer in turn makes possible a coherent interpretation of not only P3 taken as a whole in isolation, but also in conjunction with P2. As may be recalled, P2 and P3 constitute a continuous narrative progression with a single thematic thread running through (section 2.1). The common theme underlying P2 and P3 is evidently the rise (from P2 through P3.2a) and fall of Lucifer (from P3.2b through P3), or put in finer terms, his original subordination to (P2) and his subsequent rebellion against God (P3.1 and P3.2a), and his final defeat and fall (P3.2b and P3). Thematically, then, there is no privileged place for Michael (or the angel closest to God).

66 

 3 The Fall of the Rebel Angels: P3

From a structural/compositional perspective, the identification of the central figure in P3.2a as Lucifer (rather than Michael) would be in keeping with the binary opposition between God and Lucifer/Satan. The center of each register and subregister is occupied by God and/or his polar opposite, Lucifer/Satan: God/Lucifer (P2; on the specific configuration of God and Lucifer, see section 2.1 above); God (P3.2b); Lucifer/Satan (P3.1, P3.2a, and P3.3). To identify the focal figure in P3.2a as Michael—God’ agent rather than God himself—accordingly would undermine the otherwise perfect symmetry in arrangement. It should be observed further that Lucifer is represented in increased size relative to God and other angels in P3.1 and P3.2, in sharp contrast to P2, in which he is presented as a small child. The growing size may well be regarded as an icon of Lucifer’s increasing pride and arrogance. The identification with Lucifer has its plausibility enhanced by the transformation that palms undergo in hell. Provided that flails, as depicted in P17 (cf. Gollancz 1927: xli; Broderick 1978: 194, 197; Ohlgren 1992: 91), are the diabolic transforms of palms in hell, it may be inferred by extrapolation that P3.2a may well be viewed as a representation of Lucifer and his followers, a scene in which palms, heavenly preforms of flails, are manifestly in possession of Lucifer and his adherents. The prevalent, if not universal, view of P3.2b (Gollancz 1927: xli; Raw 1953: 28, 66; Broderick 1978: 92; Ohlgren 1992: 89; Karkov 2001: 50) is that God, standing at the center and attended by angels on either side, holds up three javelins aiming downward. What is at issue here is not so much the identity of the central figure—which should be none other than the Deity—as that of “the attending angels” depicted at right and left of God. While these angels are all regarded as God’s adherents in the usual reading as mentioned above, I would like to contest this standard conceptualization. Inseparably in unity with the identity of the angels is the target God aims the javelins at. Indeed, it should obviously be the falling Lucifer in the usual view. This is not, however, a self-evidently available, sole identification. As I would propose, P3.2b depicts a confrontation scene between God and Lucifer, consequent on the latter’s assertion as a god, as pictured immediately above (P3.2a). More specifically, the three angels to the left of God are rebellious ones, the rightmost one among them standing to the immediate left of God being Lucifer himself. Thus, God is facing these rebel angels, raising his javelins and aiming particularly at Lucifer. Behind the Deity stand two supporting angels. The confrontation thus involves three versus three heavenly figures, with the victorious God at the center. It may be interesting to point out that God in triumph has displaced Lucifer—who occupied the center up to this point—to the periphery on the horizontal axis (P3.2b), and eventually down into hell on the vertical counterpart (P3.3). This novel interpretation is supported on both synchronic–structural and diachronic–comparative grounds. Here, for expository purposes, I confine my argument to the former dimension and postpone the remainder till a subsequent section (3.2). The differing gestures, particularly the hands’ postures, between the left and right angels, notably the two closest to God on either side, seem to suggest that they are antagonistic

3.1 A synchronic–structural perspective 

 67

in terms of their loyalty to God: while the right angels are offensive in concert with the Lord, the left ones stand on the defensive side, even as if to beg for mercy. Of particular interest in this connection is the contrast in appearance of palms between right and left. While the one held by the rightmost angel is almost upright and lively, comparable to those depicted in the panel immediately above (P3.2a), the two palms carried by the angels on the left side look darkened and withered. These palms may thus be viewed as reflections of their bearers’ situations, doomed to prevail or be lost. In P3.2b, the angel to the left of God appears to be represented more prominently than the others. It may be surmised accordingly that this angel, as well as God, are privileged figures in P3.2b. This does not apply to the corresponding angel in P3.2a, in which the central angel exclusively is brought into focus.46 The demonstrable identity of the central figure in P3.2a as Lucifer, in turn, provides a new perspective for interpreting P3.2b. The outstanding similarity of the central figure in P3.2a to the one to the immediate left of God in P3.2b—particularly the design of the diadems they wear—may add to the plausibility of identifying the latter as Lucifer.47 Further corroboration comes from what may be characterized as the jointed border between P3.2a and P3.2b, a unique feature totally unattested elsewhere.48 An overall resemblance in shape and location would lead one to suspect that this entity corresponds to the top of the arch above the new throne in P3.1: this molded border may well be conceptualized also as the ceiling for the scene depicted immediately below.49 Such identification in turn may situate the ongoing activity firmly in Lucifer’s court. Hence, it would be but a small step to extrapolating that God and his cohorts have just marched into the palace to punish the rebels, the very scene that I interpret to be represented in the picture under consideration. Noteworthy further is another compositional parallelism, this time associating P3.2b with P2. In both, God confronts Lucifer—whether belligerently or not—standing

46 In similar fashion, the angel in the corresponding position in P16.1 is privileged over the others: he is the only angel who holds a palm. 47 The identification of the angel standing to the left of God in P3.2b as Lucifer has its plausibility enhanced by the corresponding figure’s privileged treatment in a variant representation in P16.1. In it, the corresponding angel exclusively wears a diadem and holds a palm among the six angels represented. The monopoly of these symbols by this angel suggests his privileged status. Thus, it may follow that this angel be identified as Lucifer. In this respect, Gollancz’s interpretation must be criticized as inconsistent and incoherent: while he correctly characterizes the angel confronting God in P16.1 as rebellious (Gollancz 1927: xli), he (Gollancz 1927: xxxix) apparently regards the corresponding figure in P3.2b as obedient. 48 It may be of interest to draw attention to a similarity to part of decorated capitals, specifically a long staff of a letter like Þ as on page 73. See also note 55 below. 49 The unique decorated boundary takes on greater thematic significance in a broader context, as observed earlier. Above this outstanding border is the rise of Lucifer culminating in his receiving the symbols of victory in P3.2a, and below it is his fall, starting with his defeat by the hand of God (P3.2b) and ending in his perpetual bondage in hell.

68 

 3 The Fall of the Rebel Angels: P3

to his left. It should be noted that the figures to be identified as Lucifer in P2 and P3.2b are posed quite alike. No less significant is the overall similarity in God’s posture in his confrontation with the putative Lucifer standing to the left. The two pictures may therefore be characterized as variants of the same underlying motif. This common conceptual basis may be adduced as an additional piece of evidence for identifying the figure in question in P3.2b as Lucifer, drawing on the parallel configuration in P2, for which the identity as Lucifer is far more transparent.50 In contrast to the earlier two registers, the third one, P3.3, poses no immediate difficulty in its overall interpretation. In the upper half is represented the falling Lucifer, along with his company and parts of his throne. The lower half depicts Lucifer fallen to the mouth of hell, turned into Satan, and bound with chains (Gollancz 1927: xxxix; Broderick 1978: 93; Ohlgren 1992: 89). A greater significance of P3.3 lies in its variation in conjunction with P3.2b: P3.3, coupled with P3.2b, is partially replicated in two other variant representations, P16 (Pl. 10) and P17 (Pl. 11). In correspondence to P3.2b/P3.3, the upper halves of P16 and P17, referred to as P16.1 and P17.1, respectively, depict the event/state in heaven, while the lower half, P16.2 and P17.2, represent the scene in hell. These pictorial variations depict differing phases of the same event—the defeat and fall of Lucifer and his company— and resultant state of affairs in terms of temporal sequencing. Inside heaven, the three relevant pictures may be ordered as follows: P16.1 > P3.2b > P17.1. First, God, supported by his followers, reproaches Lucifer and his cohort (P16.1); then, God attacks his enemies (P3.2b); and finally, with the enemies thrown out of heaven, God is seated serenely, attended by seraphim (P17.1). As for the scenes outside heaven, they unfold in the following order: P3.3 > P16.2 > P17.2. That is, Lucifer and his adherents are falling from heaven, Lucifer turning Satan and bound at the mouth of hell (P3.3); partly overlapping with P3.3, some rebel angels are still falling from heaven toward the hell mouth, while others—including Lucifer—have already passed through it to the bottom of hell (P16.2); eventually, having all fallen from heaven, the defeated angels are now enclosed within the walled hell (P17.2). Put altogether, the six scenes may be sequenced from beginning to end as follows: P16.1 > P3.2b > P3.3 > P16.2 > P17.1/P17.2. That is, starting with God’s rebuke (P16.1), followed by his strike (P3.2b), the rebels are falling from heaven to the mouth of hell (P3.3), have been swallowed into hell (P16.2), and remain enclosed in hell (P17.2), whereas God is seated at peace in heaven (P17.1). Far from mechanical repetitions of the preceding images (e.g., P3.3) as implied by Doane (1991: 41), these variant pictures are carefully designed and integrated into a unified set through systematic complementation and differentiation.

50 P2—two-tiered—may be regarded as an elaborate version of P3.2a and P3.2b, both being of a single tier.

3.2 A diachronic–comparative perspective 

 69

3.2 A diachronic–comparative perspective In terms of the overall partition of a single page into three major registers for pictorial representation, P3 most closely resembles Vivian 3v (Pl. 26) and 386v (Pl. 31), and their cognates San Paolo 3v (Pl. 33) and 310v (Pl. 37). The further division of each layer into two subregisters, however, is distinctively characteristic of P3. Moreover, the use of straight bars as boundaries between the registers distinguishes Vivian 3v, 386v, and San Paolo 3v from San Paolo 310v, which is demarcated with a row of cloud-like white puffs for the most part and a double-layered floor for a limited section. Of particular interest, in P3, too, a similar-shaped double-layered floor is used as a secondary border, the minor boundary that subdivides the right half of the top register (P3.1). Thus, as far as the three-layered articulation of space in general is concerned, these four pages in the Vivian and San Paolo Bibles are exactly the same as P3. As regards the manner in which the segmentation is marked out, they exhibit a complementary similarity to the Junius page, with San Paolo 310v (double-layered floor) differentiated from the other three (solid bands). That is, the framing of P3 is executed by combining the markers characteristic of Vivian 3v/386v/San Paolo 3v on the one hand and San Paolo 310v on the other. From top to bottom, the registers are sequenced in narrative development in all of the pages at issue. Within each register, however, temporal sequencing does not obtain in the same way. With respect to P3, only the top register is internally differentiated in temporal terms, progressing from right to left (the upper half) and then from the upper to the lower half. The remaining two registers are vertically oriented from upper to lower subregisters, without any horizontal movement in play. As for the two Bibles, the top register is organized horizontally from left to right, a reverse directionality of the Junius page. Much the same applies to the middle registers of Vivian 386v and San Paolo 3v, and the third registers of San Paolo 3v and 310v. The remainder —with the exception of the second register of San Paolo 310v—namely, Vivian 3v2, 3v3, and 386v3, are devoted to a single scene, manifesting no sequential development within a register. San Paolo 310v2 is exceptionally complicated in its internal progression: the three scenes represented there start from the left side, move then to the rightmost, and finally end in the middle section. In terms of its overall framing and internal dynamics, P3.1 may thus be characterized as broadly isomorphic to and, by diachronic implication, inspired by Vivian 3v, Vivian 386v, San Paolo 3v, and San Paolo 310v in different ways. A fundamental question arises: Why is it that these pages, rather than other illustrations in the same works, came to serve as artistic sources for the illumination of P3 in particular? To address this question, one needs to go beyond pure formalism and delve into inner substance of the putative models. In other words, the parallel/isomorphic organizations that have been brought to light above must be substantiated by offering a principled account of the conceptual basis for and mechanism of adopting and adapting the model illustrations for the designing of the Junius page.

70 

 3 The Fall of the Rebel Angels: P3

Before proceeding further, it may be appropriate here to provide an overview of what the four model illustrations represent, by drawing on Kessler (1977: 84–87, 111– 119). On the one hand, Vivian 3v and San Paolo 3v, appearing as frontispieces of Jerome’s letter to Paulinus, illustrate Jerome’s activities, ranging from his departure from Rome to his distribution of copies of the Vulgate Bible he had completed. More specifically, the first registers of both Bibles are allotted to Jerome’s leaving Rome and his offering coins to his Hebrew teacher in the school building. In the second register of Vivian 3v, Jerome, the central figure, explicates the text to four women on the left, while three monks sitting behind Jerome on the right side of the bench write down his explanations. By contrast, the corresponding panel of San Paolo 3v consists of two different episodes: in the left half Jerome argues against two monks standing on the left, and in the other half he interprets for two seated women while a monk and another woman record his teaching. In the final registers of both works, Jerome is shown delivering copies of the Vulgate. In San Paolo alone, however, this episode is preceded by another scene in which Jerome dictates to four monks, two before him and two behind. On the other hand, Vivian 386v and San Paolo 310v, both designed as frontispieces to the Pauline Epistles, visualize the conversion of St. Paul. The first register of Vivian represents Saul’s departure from Jerusalem (left section), then his blinding and fall (middle section), and finally his entry into Damascus (right section). In San Paolo 310v, in contrast, the second and third episodes are illustrated, respectively, in the right side of the first register and the far left of the second register. More critically, instead of his departure scene, Saul is shown receiving letters from high priests in the left side of the first register in the San Paolo Bible. Apart from the different episodes being thus rendered in the opening of the first register (Saul’s departure from Rome in Vivian versus his receiving letters from priests in San Paolo), the two Bibles differ in the number of Saul’s attendants as recorded in the final episode of his entry into Damascus, one in Vivian and three in San Paolo. Similarly, the choice of differing scenes for the first part of the initial register between Vivian (Saul’s departure from Rome) and San Paolo (Saul’s receiving letters) results in the varying size of the companies depicted, two (Saul and his attendant in Vivian) and seven (five priests in addition to Saul and his companion in San Paolo). Despite the divergent numbers of attendants represented in the third scene, Saul’s blinding and falling to the ground is witnessed by two disciples of his in the Vivian and San Paolo Bibles. The reuse of the Paul and Jerome illustration pages for representation of Lucifer’s rise and fall, as will be fully substantiated below, may command initial plausibility. Their choice seems to have been motivated on the whole on partly complementary and partly mutually reinforcing grounds. The conversion of Paul would have been chosen as a model primarily for conceptual reasons, namely, due to the central motif appreciated to be analogously involved. At issue is God’s punishment and Saul’s resultant fall to the ground, which is strongly reminiscent of Lucifer’s fate. By contrast, the Jerome cycle has a locational advantage. It occupies the privileged place of constituting the opening picture of a volume or one closest to it. In Vivian, the page makes a frontispiece to the

3.2 A diachronic–comparative perspective 

 71

whole volume, while in San Paolo it is preceded solely by a dedication page (1r). In this light, the Jerome cycle of pictures would have more likely attracted the artist’s attention than any other illustrations in the available models when he engaged with illuminating a comparably localized page. Moreover, the Paul and Jerome pages are associated with each other inherently in narrative terms, as in other respects (cf. Dutton/Kessler 1997: 58): the journey (with points of departure and arrival represented) and preaching (with addressees specified) episodes, progressed in this order, figure prominently in both. Last but not least, the lack of intrinsic association with the text of Genesis would have made these illustrated pages redundant and hence accessible as most convenient resources for reuse and adaptation as visualizations of Genesis episodes. Having set the necessary background, I resume comparative and genealogical inquiries by exploring major constituents of each scene in P3 as to their sources, cognitive bases of analogical adaptation, and structural mechanisms of reconfiguration. Starting with P3.1, one may characterize the entry into the new palace represented in the upper half as conceptually analogous to Jerome’s departure from Rome and his visit to his teacher in Vivian 3v1 and San Paolo 3v1 on the one hand, and Saul’s departure from Jerusalem and his entry into Damascus depicted in Vivian 386v1 and San Paolo 310v1/310v2, although the physical movements involved are presented in opposite directions, going from left to right (Vivian and San Paolo) versus from right to left (Junius).51,52 More specifically, in the initial scene of P3.1 (the upper right), there is a procession of four angels depicted. Bearing in mind that P3.1 has its linear progression reversed in contrast to the Vivian and San Paolo predecessors, one might be led to expect that this group of angels evinces some affinity with the functionally equivalent events recorded at the left in the two Bibles. In actuality, however, no more than a single man (Vivian 3v) or two (Vivian 386v1 and San Paolo 3v1) are represented in the departing scene on the left side, whereas as many as seven appear in the corresponding initial episode in the left half of San Paolo 310v1. One accordingly would be hard pressed to account for the numerical disparity in evidence. Alternatively, however, one may link images between source and recipient by proceeding frame by frame in the same horizontal direction in disregard of the reverse directionality of narrative development involved. Consequently, the upper right image component of P3.1—the first narrative segment—is matched with the entry scenes—the last narrative segment—of the presumable Touronian models. As it turns out, a significant correspondence comes to light between the narrative beginning in P3.1 (on the right) and the last—the entry into Damascus—of the three51 As noted above, the episode visualized in the left section of San Paolo 310v1 is an event that took place prior to Saul’s actual departure, namely, his receiving letters from high priests. Of central relevance here, however, is that Saul’s blinding occurred during his journey from Jerusalem to Damascus, the succession of events that is commonly presupposed in both Bibles. 52 As may be recalled, San Paolo 310v deviates from the other illustrations by allocating the goal scene to the initial section of the second register.

72 

 3 The Fall of the Rebel Angels: P3

part journey scenes in San Paolo 310v2 (on the left). Exactly four people including Saul are represented in the entry-into-Damascus scene in San Paolo (310v2). In corresponding fashion, four angels are found also in the equivalent position in P3.1 (the upper right), recategorized now for comparison as the initial rather than the final phase of the journey scenes. As discussed below, Lucifer may most plausibly be derived from the pair of Paul’s attendants standing in the middle section of Vivian 386v1 or in the right half of San Paolo 310v1. Accordingly, Lucifer is not included in the body of four angels under consideration in P3.1. In other words, Saul in the entry-into-Damascus scene seems unlikely to have been reused as Lucifer in P3.1 in terms of overall configuration. Specifically, the preform of Lucifer at the center of P3.1 belongs to a separate domain of representation, namely, the second phase, rather than the last, of the set of journey scenes (Vivian 386v1 [middle] and San Paolo 310v1 [right]). Therefore, when Lucifer joins the company as is probably the case with the behind-the-palace scene, the entire group of angels amounts to five, as depicted in the upper left of P3.1. In this connection, the four angels in the lower right section deserve special treatment, depending on the alternative readings available, as pointed out partly in the previous section. First, the lower right scene may be conceptualized as independent of and subsequent to the upper one. Specifically, the angels below may be interpreted as offering crowns at the coronation that is to take place within the new palace. In that event, two further alternative interpretations become available, hinging on the identities of the angels, that is, whether the four angelic beings in the two tiers are the same four. If in the affirmative, then it must follow as a matter of course that these four beings appear below as well as above the boundary. If in the negative, one may well hold formal parallelism as responsible for the resulting agreement in numbers. In either event, the appearance of four angels, rather than any other number, may be understood as a duplication of the corresponding procession in the above tier, whether by virtue of the real identities of the angels involved in the two layers or simply for the sake of structural harmonization. Now that the number of angels depicted and their distribution in the three subregisters in P3.1—the upper and the lower right, and the upper left—have been shown to be derivable from San Paolo 310v, it is appropriate at this point to offer a derivational trajectory for Lucifer standing at the center of P3.1. Two complementary sources may be identified as underlying the figure in question. First, as alluded to above, the centrality and the particular pose they assume reasonably point to the two attendants of Saul, depicted in Vivian 386v1 and San Paolo 310v1. In either version, Saul’s guards are shown raising one hand and looking toward their master in astonishment at and for protection from the beam of light emitted from the sky. Depicted at the center and in prominent form—doubly figured, with a unique posture and in largest size (notably in San Paolo 310v1), the two men would have served as a compositional model for Lucifer in P3.1, reconfigured in doubled size and overwhelming the other figures. In this way, the two astounded attendants in the center—particularly in San Paolo 310v1—are reconfigured to be a representation of Lucifer. Corresponding to the double figures in the model and

3.2 A diachronic–comparative perspective 

 73

the two-storied building (with a focus on the throne on the upper floor; see below), Lucifer, a single figure, undergoes expansion along the vertical axis. Of no less significance, the opening scene of Jerome’s departure from Rome, provided with his beckoning motion to the ship (Vivian) or his readiness to board it (San Paolo), would have motivated the artist towards analogical adaptation by virtue of the common conceptual bases involved. Concretely, Jerome beckoning toward the ship (Vivian 3v1) would have been transformed into Lucifer pointing to the throne (P3.1). Noteworthy here is the overall similarity in shape between the ship and the throne. Also worth pointing out is that Jerome holds a staff, as does Lucifer. The presumable hanging of curtains from the arch above the new throne (P3.1) and the sail of the ship would have facilitated a metaphorical transference by virtue of their similar color, shape, material, and location relative to the seats. This metaphor would have come to mind readily, given the cooccurrence of sails and curtains manifested in the first register of San Paolo 3v1 (center and right). Of related interest, the depiction of Lucifer (the fifth angel counted from the left) as being about to enter the palace in P3.1 appears to be parallel to Jerome’s readiness to board the ship (San Paolo 3v1) on which four men—analogous to the four angels on the left in P3.1—are already on board. Reversing horizontally this configuration of the five men results in a virtual reproduction of the scene on the upper left of P3.1, thereby lending substantial support to the interpretation, proposed in the preceding section, that the four angels are prepared in the building to welcome Lucifer. The only notable difference concerns the exact arrangement of the three men/angels facing Jerome/ Lucifer—directly in front of Jerome or intervened by another figure between the three angels and Lucifer. This compositional disparity, however, is subject to an explanatory account with reference to San Paolo 310v1 (left), as treated below. Thus, the figure of Jerome in the departure scene of the two Bibles would have undergone redrawing and reconfiguration so that his original identity was split into two separate figurations of Lucifer, the entering one (derived from San Paolo 3v) and the pointing one (derived from Vivian 3v). This double use was made possible by the availability of two complementary representations in the models. In the preceding section, I claimed that the central figure of Lucifer in P3.1 belongs to two distinct scenes by conflation and plays double roles accordingly. On the one hand, he summons his followers to his new palace, and on the other he is receiving crowns in the palace. Such a double representation is demonstrably at work in one of the models, namely, Vivian 386v1. At stake is the fallen Saul at the center of Vivian 386v1. This figure condenses two scenes—his reaction to having been struck by the light while walking,53 and the consequent fall to the ground. The raised right arm corresponds to the first scene, whereas the remainder belongs to the situation after the fall. All pictorial unnatu53 The image is immediately obtained by rotating counterclockwise the fallen Saul to upright position. While Kessler’s (1977: fig. 175) reconstruction, minimally distinct from Paolo 310v1 (right), is meant to be diachronically real, my proposal is intended to be a synchronic virtual reality in the mind of the artist and sympathetic viewers.

74 

 3 The Fall of the Rebel Angels: P3

ralness and awkwardness attending on this figure and surrounding ones may be readily understood as consequences of this marked operation. Given that the implied walking figure struck by the light was denied its own realization by subsumption under the fallen one, the light had to be redirected to the departure scene for materialization. The other option of relocating to the fallen figure would be less natural in terms of cause and effect. Since at the moment of departure Saul was not yet attacked in actuality, he duly showed no sign of disturbance, despite the apparent attack as depicted by way of compromise. This reading is no less capable of accounting for the otherwise puzzling depiction of the two attendants at the center: Why are they walking far ahead of Saul, when they—at least one of them—had followed their leader at departure? It seems contradictory that Saul is struck by the light at his departure whereas his attendants were shown on their way to Damascus, looking far back to Saul’s company presumably including one of themselves. A further puzzle is that the attendants, astonished and turning back, exhibited no apparent concern for the fallen master. These bizarre features all succumb to understanding upon acknowledging the implied presence of a walking figure of Saul corresponding to the fallen one just behind the two attendants. On their way to Damascus, the company was hit by the light, his followers turned to Saul in fear and astonishment, and thereafter he fell to the ground. Thus, P3.1’s indebtedness to the Touronian Bibles goes beyond individual figures and compositions so as to penetrate into a profound and abstract level of pictorial organization. Moving on to the group of four angels apparently behind but most plausibly inside the building in P3.1 (excluding the relatively large fifth figure, supposedly Lucifer), they are strikingly similar, in terms of overall configuration and orientation, to the men represented in corresponding position in San Paolo 310v1—the high priests (excluding the one directly facing Saul to the right, who finds no equivalent in P3.1, probably due to his being the least visible in the original scene). Of paramount interest, the first three of the four angels at the far left are oriented to the right, the fourth relatively small one alone is facing leftward, exactly the same configuration of the four leftmost men in San Paolo 310v. Further, the fifth angel, relatively large in size, to the immediate left of Lucifer at the center of P3.1 is conspicuously comparable to Saul together with his guard to his right, both standing at a short distance from the high priests, just as the fifth angel seems to keep some distance from the gathering of the four angels to the left. These compositional agreements, down to the details, can scarcely be explained without postulating a genealogical affinity between P3 and San Paolo 310v, and more specifically, positing the latter as a derivationally earlier model. Lucifer in P3.1 is thus doubly composed, each being a composite figure derived from the multiple characters appearing in the two distinct type scenes represented in the two cognate manuscripts: the central figure of Lucifer is a reconfiguration of the two astonished attendants in Vivian 386v/San Paolo 310v and the beckoning Jerome in Vivian 3v; the other figure immediately to the upper left of Lucifer at the center owes to the boarding Jerome in San Paolo 3v and Saul and his attendant facing the high priests in San Paolo 310v.

3.2 A diachronic–comparative perspective 

 75

The two-storied palace on the left in P3.1 is equipped with a throne on the second floor. The arrangement of the throne on the upper level is obviously designed to emphasize a higher place privileged for Lucifer in accordance with the textual description, as observed above. The setting of the two-tiered building in turn requires for pictorial harmony a double-layered panel for representation on the facing half of the page. Thus, the two processions of angels as discussed earlier are necessitated on compositional grounds as well. The hall in itself is modeled specifically after the buildings in Vivian 386v1, 386v2 (left), and Vivian 3v1 (right), whereby the schoolhouse (Vivian 3v1) serves as an overall frame structure for Lucifer’s palace by flanking it with towers attached with two layers of roofs. In the original picture, the central part of the schoolhouse, apparently single storied, is left hollow presumably so as to depict the two major figures in conspicuous form. In the new context, however, the building needs to be pictorially filled in to accommodate the solid two-storied structure, as conceptually dictated (see above). In so doing, the original roof has to be raised above the side towers, as realized in P3.1. At the lower level of the main part, the arched gate provided with two square windows (or turrets) above it may be derived from the analogously shaped city gate with turrets in corresponding position, which is found on the left corner of Vivian 386v1. Although standing on its own in the original, the Junius gate is incorporated as the first floor of the palace and expanded accordingly so that it is flanked by two-storied towers, added with a clerestory and topped with a roof, for each level. The construction is largely in agreement with the side towers of the schoolhouse in Vivian 3v1, which plausibly would have served as a model. A notable difference, however, is that the roofs on the top of both towers are tented—in contrast to gabled (as in Vivian 3v1)—in form, which is noteworthy, as such roofs seem to be unattested in Vivian and San Paolo.54 As for the upper level of the palace, Ananias’s house (Vivian 386v2) with an arch with a circular hollow on either of the upper sides constitutes a striking counterpart to the Junius palace. Both buildings and also the schoolhouse (Vivian 3v1) are represented in side-gabled view, thereby suggesting their close affinities. God’s hand from above emitting rays of light to Ananias will immediately conjure up the image of a royal palace that is analogously attended from heaven, such as is represented in Vivian 423r (Pl. 32). This conceptual link may in turn motivate appropriation of Ananias’s house as Lucifer’s court. The usual attending decorations like curtains (Vivian 423r and San Paolo 3v1),55 54 Conical roofs are depicted in Vivian 3v2, San Paolo 170v, and 295v, however. 55 The enigmatic decorations applied to the arch above the throne would seem to be beyond a precise identification in synchronic terms. Admittedly, they hardly fit natural appearances of curtains, despite my suggestion proposed in the text from a diachronic–comparative perspective. Another plausible interpretation would be vegetative or zoomorphic decorations, very much comparable to those executed on capitals. As a close parallel, one may adduce a decorated Ð on page 43. The competing readings are far from mutually exclusive particularly in the diachronic dimension. I would speculate that the original representation of curtains based on the model might have been reinterpreted as ornaments attached to the arch in the course of transmission.

76 

 3 The Fall of the Rebel Angels: P3

lacking in Ananias’s simple dwelling, may be readily supplied to attain a regal dignity as needed. Most importantly, Ananias’s bed (Vivian 386v2) in itself may naturally evoke association with the throne situated on the second floor in compositional and metaphorical terms, specifically because of their overall comparable shape and the common presence of cushions or mattresses. Accordingly, the formal and functional similarity would have inspired the artist to adapt Ananias’s house as the second floor of the hall. In this regard, the design of the throne requires comment. Being elongated and having no back, it looks more like a bed (Vivian 386v2 and San Paolo 310v2) or a cushioned bench (Vivian 3v2). Further, curled inward, the scroll elements on the sides—the tips of the armrests corresponding to the head- and footboard legs of a bed—are the reverse direction from what is usually the case—coiled outward, as in San Paolo 3v2. In any event, the peculiar shape of the throne in P3.1 becomes amenable, in part if not in whole, to explanation by postulating Ananias’s bed in Vivian 386v2 as its underlying preform. On the whole, then, Lucifer’s hall in P3.1 is constructed by integrating elements drawn from the buildings depicted in Vivian 3v1 (right, schoolhouse), 386v.1 (left, the city gate of Jerusalem), and 386v.2 (left, Ananias’s house). The combinations were executed carefully according to cognitively motivated designs. Ananias’s bedroom is incorporated as the center of the hall, the upper level housing the throne, commanding a higher position in heaven. Through adaptation and elaboration, the originally humble construction is embellished on the arch above—hangings—and the throne below—side attachments—to add grandeur. The necessary lower floor owes its basic form to the city gate (Vivian 386v1). In order to give the whole building a structural integrity and a regal majesty, towers are attached to its sides, the outer framework that is conveniently supplied by the schoolhouse (Vivian 3v1) because of its hollow center, which can readily afford to accommodate the solid bodies derived from the two separate constructions, as discussed above. It is interesting to note here that the schoolhouse and Ananias’s home are almost identical in form with respect to their long gabled roofs, the structural correspondence that provides a formal foundation for the integration of the two building units. Broderick (1978: 96–102; 1983: 167–168) argues that P3.1 exhibits conspicuous correspondences with Utrecht 26r (Pl. 75; DeWald 1932: 22; Utrecht Psalter 2015: 162–163), so much so that the former must be regarded as depending on the latter. His claim is grounded on two major arguments that he adduces. At issue first are the two rows of angels, four in each, on the right of P3.1—the upper being empty handed and the lower carrying crowns—in comparison with the two groups of gift-bearing people, females above and males below on the right side of Utrecht 26r. Struck by their close similarity “in pose and function,” especially the correspondence in gesture between the empty-handed angels and the gift-bearing female figures, Broderick (1978: 97–98) is led to posit a genealogical relation between the two drawings. Second, and in a similar vein to the first observation, Broderick (1978: 98) draws attention to a parallelism between the four angels behind the palace on the left of P3.1 and the three figures behind the throne

3.2 A diachronic–comparative perspective 

 77

in Utrecht 26r. Standing behind the empty throne, both groups are alleged to be similarly configured and “[e]ven more telling[ly]” so at that than the first comparable instances. While the resemblances Broderick points out seem too remarkable to be dismissed outright, there are notable differences that may preclude one from attributing them confidently to the close genealogical relation that he claims. Particularly forbidding seems the second observation above. The putative parallelism is not as compelling as Broderick would have one believe. For one thing, the two groups of figures involved are situated distinctly relative to the throne and the palace: immediately behind the throne in Utrecht, but not remotely so in Junius; in front of the palace in Utrecht, but at the back of it in Junius. In this light, the configurations concerned must be regarded as differing rather markedly between the two drawings. For another, the numbers of figures depicted do not match, with four in Junius as opposed to three in Utrecht. Moreover, neither the thrones nor the palaces exhibit interesting resemblances between them (for details, see below). In comparison, the double rows of gift bearers apparently constitute a closer counterpart to the corresponding processions of angels in general profile. In addition, the same numbers (four) of figures are represented in the lower groups carrying crowns in P3.1 and Utrecht 26r. Indeed, the crowns are hardly matched in their details, but the distinctions involving small artifacts such as the ones at issue here obviously carry less weight in determining a genealogical relation than larger objects such as architectural structures in the earlier case. All in all, then, the parallelism pertaining to the double progressions of figures might be accepted as genuine. Placed in a larger context, the above reasoning will prove to be inappropriate. There turns out to be a more outstanding parallel in evidence, which may discredit the above hypothesis. At issue is the top group of five female figures to the right of the throne in Utrecht 26r, with the leftmost one standing at a slight distance from the four crowded together to the right. These five figures altogether are matched more closely with the upper row of four angels plus Lucifer—five in total—than is the second row of females (six in all) that Broderick claims. The configurations involved are outstandingly parallel: in both Junius and Utrecht, the leftmost figure keeps some distance from and glances back to the remaining four to the right. Moreover, the latter four are empty handed in Utrecht, precisely as in Junius, and they are correspondingly postured, including hand gesture, with a minor difference that the third figure from the right faces in the opposite direction in Utrecht, but not in Junius. In this way, that exclusively the lower angels carry gifts in P3.1 is explained only by positing the upper two rows of people in Utrecht 26r as a source for Junius. In contrast, should one be committed to Broderick’s proposal that the upper row of angels is modeled after the middle group of gift-bearers, it remains to be explained why the corresponding angels in P3.1 have nothing in their hands to offer.56

56 Even in the second row in Utrecht 26r, as in the top row, there is one figure (second from the extreme right) oriented differently from the remainder. In this light, the exceptional feature observed in the text hardly counts as a decisive piece of counterevidence.

78 

 3 The Fall of the Rebel Angels: P3

Thus, the first two rows of females should more convincingly be identified as a model for the composition of Lucifer and the two groups of angels to the right than the second and third rows of figures, as Broderick (1978) believes. Added to the formal considerations explored above, there remains a conceptual dimension to be brought to bear on the issue of deciding between the two competing candidates for the Utrecht source of Junius P3.1. While the middle and lower groups of figures visualize Psalm 44.13 (45.12), “The people of Tyre will seek your favor with gifts, the richest of the people with all kinds of wealth” (Coogan 2010), the upper group constitutes a visual representation of Psalm 45:14–15, “. . . behind her [= the princess] the virgins, her companions, follow./ With joy and gladness they are led along as they enter the palace of the king” (Coogan 2010). This content suits optimally what P3.1 most likely would have been intended to convey: Lucifer leads his supporters to his new palace, as interpreted by drawing on impressionistic readings as well as textual (Genesis B) and pictorial (Vivian/San Paolo) sources. Accordingly, the princess, the leftmost female in Utrecht, is reshaped as Lucifer through adaptation, while the four virgins are reconfigured as his followers. The conceptual independence of the top two groups in Utrecht 26r also accounts for that lack of unity, or their cognitive disjunction, in pictorial terms, in P3.1. The absence of meaningful coordination between the two rows in question may therefore be regarded as an inheritance from the model drawing, in which the two groups are simply juxtaposed without any intrinsic connection with each other. As argued in the foregoing discussion, the configuration of Lucifer and the two groups of angels to the right most plausibly would have been modeled after the upper two rows of female figures in Utrecht 26r. At this point, a fundamental question arises: How does this conclusion relate to the claim presented earlier that P3.1 in general was inspired by Vivian 3v/386v and San Paolo 3v/310v? Settling this issue requires a substantive assessment of the two competing accounts along a set of empirical criteria. More specifically, in making such an empirical evaluation, it is of vital importance to compare how successfully and convincingly these alleged models can provide diachronic–genealogical explanations for significant properties of P3.1 As should be clear from the discussion so far, P3.1 is divided, in terms of its structural organization, into two zones, the right and the left side, which are in turn composed of two constituents each, upper and lower. The right half is devoted to the representation of Lucifer and the two processions of angels. This major portion consists of two subregisters—the upper and the lower group of angels, respectively. The two vertical subdivisions intersect by virtue of the overlapping figure of Lucifer. The left half of P3.1 depicts the group of five angels behind the palace, which naturally subdivides into the angels and the palace even in the absence of an explicit boundary comparable to the right half. As with the vertical dimension between upper and lower, the central figure of Lucifer serves as a joint for connecting the two sides of the panel along the horizontal axis. It accordingly will be most efficient to make a comparative assessment of the two models with reference to these (sub)divisions.

3.2 A diachronic–comparative perspective 

 79

Starting with the right side, the double-tiered composition in itself may be readily explained as a direct consequence of the original, two-layered organization in Utrecht 26r, as is the left-facing orientation. By contrast, an appeal to Vivian and/or San Paolo as alternative sources of inspiration would require explanation in a less straightforward, apparently more roundabout way by addressing a deeper motivation of redesigning and consequential multi-stepped executions. As may be recalled, the ultimate need to accommodate a throne in the upper level of a stately palace results in the two-storied building which should be about twice as tall as angelic figures, with due respect to compositional harmony. Thus, the aim to keep compositional balance would have favored the correspondingly two-layered arrangement of angels. In similar fashion, the artist’s pronounced preference for depicting a left side view of a central figure must be invoked to account for the reversal in face orientation as well as direction of movement in Junius as opposed to its putative models, Vivian/San Paolo, as will be discussed in regard to P3.2b below. At first glance, then, Utrecht 26r appears to be superior as an explanatory account, as it can offer an immediate path of derivation from original to copy, whereas Vivian/ San Paolo requires a long chain of intermediary processing to reach the final output. On deeper analysis, however, it turns out that the positing of Utrecht 26r as a model for P3.1 is in need of the same initial motivation as is postulated for its Vivian/San Paolo counterpart. In the absence of such a conceptual motive, it would be hardly intelligible why the artist was inspired at all by Utrecht 26r in general and the upper two groups of women in it in particular in illustrating part of P3.1 in Junius. Or posed differently, why was he not content with adopting a single procession of figures or obeying the Utrecht model most faithfully to reproduce maximally the three groups of followers? Thus, one is thrown back to where one started with the Vivian/San Paolo sources: the original vision of Lucifer creating a two-storied palace for his occupation. It was this specifically articulated mindset that shaped the artist’s subsequent executions. In the final analysis, both accounts must be predicated on the same compositional motivation of accommodating a throne on an upper level of the palace, the second floor in the minimally differentiated two-storied building. Given the underlying motive that must commonly and eventually be invoked for the reuse of Vivian 3v, 386v/San Paolo 3v, 310v on the one hand and Utrecht 26r on the other, one may wonder why these specific models were chosen from available candidates for adaptation. Why were other page illustrations of the same works not selected as sources? In this respect, the Vivian/San Paolo hypothesis seems to be more explanatory and revealing than the Utrecht thesis. As argued above, the Vivian and San Paolo Bibles, specifically the pages devoted to the episodes of St. Paul’s conversion, have more profound conceptual affinities to Junius P3 than does the Utrecht counterpart. In the former two, the rise and fall of God’s enemy, and God’s punishment imposed on his opponent, are featured as central motifs, whereas these elements of disloyalty and retribution are relegated to the margin in Utrecht 26r, if they are appreciable at all. Rather, in the Utrecht drawing in question, God’s blessing on the king and his resultant pros-

80 

 3 The Fall of the Rebel Angels: P3

perity on earth are set in focus. Whatever resonance may by any chance be effected through Psalm 45.15 “With joy and gladness they are led along as they enter the palace of the king” (Coogan 2010), the Utrecht’s conceptual connection with Junius P3 must be characterized as tenuous. Thus, Vivian 386v and San Paolo 310v are qualified in their essentials for serving as comprehensive models for Junius P3 as a whole, not simply for its initial register, P3.1, while Utrecht 26r is limited to individual configurations in the range and depth of its inspiration. As for specific compositional elements in P3.1, some may well be regarded as based on Vivian/San Paolo, others on Utrecht, and still others derivable from both. First, the involvement of four angels, no more and no less, may be explained by either model. There are precisely four figures depicted in the entry scene in San Paolo 310v (the left of the second register) on the one hand, and in the top group of females in Utrecht 26r on the other. Second, the use of a double-layered floor as the half-border between the upper and lower subregisters in P3.1 receives a principled account by identifying its origin as the floor of the priests’ building in San Paolo 310v—also serving as a partial border separating the first register from the second in this San Paolo exemplar. Noteworthy in this connection is the inversion whereby the original starting point on the left side of the page is relocated to the right in Junius P3.1. To adhere to Utrecht 26r as the sole source of P3.1, however, would be incapable of explaining the use of a floorboard—to say nothing of its two-layered construction—as a division between the upper and lower rows of angels. Such an artificial, rather than naturalistic, manner of demarcation is unknown to the Utrecht Psalter in general. Third, and inseparable from the second point raised above, a double-layered foundation is also used partially as a division between P3.1 and P3.2a. This time, the demarcation is executed in the same place, the left side of the first register. The recurrence of the same construct in the two locations that are analogous in different ways to the original one in San Paolo 310v must be regarded as eloquent testimony that this page played a significant role in illuminating P3.1 in Junius 11. Fourth, that exclusively the lower angels carry gifts is explained only by positing Utrecht 26r as a source, whereas the act of gift-bearing in itself cannot be derived from Vivian/San Paolo. Fifth, the putative entry scene on the left side of P3.1 is subject to explanation by assuming the right-to-left movement through the upper half of the register, a mode of inherent dynamism which is derived from Vivian/San Paolo with a reversal in directionality. In contrast, Utrecht psalm pictures in general are each composed by an overall harmonious arrangement of individual images that are literal visualizations of words and phrases of a constituent verse (e.g., Dufrenne 1978: 195–198; McKitterick 1990: 311–312; Horst 1996: 55). The resultant whole medley of disparate images is static, whereby all visual constituents executed in a single panel, 26r for example, are independent of and unrelated to each other conceptually, defying an overall temporal/causal interpretation relative to one another.57 57 A causal relation inherent within an individual verse is another matter, such as the visualization of Psalm 45.5 “Your arrows are sharp in the heart of the king’s enemies; the peoples fall under you.”

3.2 A diachronic–comparative perspective 

 81

Sixth, the placement of a throne outside the palace at the center in P3.1 can be explained only by recourse to Utrecht 26r, not to Vivian/San Paolo, where there is no such object immediately available. Seventh, the particular ways in which the five angels behind the palace are represented in P3.1 in regard to their orientation and facial expression yield to a credible account by tracing them back to the corresponding group in the building in San Paolo 310v. Eighth, the specific construction and design of the palace is explained by an appeal to the Vivian/San Paolo model. Inasmuch as Utrecht 26r offers a face-gabled view without an arch and side towers, these architectural features characteristic of Junius P3.1 remain beyond comprehension under the hypothesis that Utrecht 26r served as an exclusive model. Ninth, in a similar vein, the shape of the throne and the design of the cushion resist explanation insofar as one subscribes to the Utrecht 26r model.58 By way of summary, the nine individual design features (1 through 9), along with the overall conceptual framework (0), are distributed among the five illuminated pages, as shown in Table 7. Table 7. Distribution of P3.1’s design features in the five putative models Feature

Vivian 3v

Vivian 386v

San Paolo 3v

San Paolo 310v

Utrecht 26r

0  1  2  3  4  5  6  7  8  9 

          √      √   

√          √      √  √ 

          √    √     

√  √  √  √    √    √     

  √      √    √       

Key: 0. Conceptual framework; 1. Four angels in the upper right group; 2. Double-layered floor between the upper and lower rows of angels; 3. Double-layered floor below the palace; 4. Exclusively the lower group of angels bearing gifts; 5. The entry scene on the left; 6. The throne placed outside the palace; 7. The arrangement of five angels behind the palace; 8. The architectural design of the palace; 9. The design of the throne.

The above distribution pattern may allow for the following generalizations. Given the overarching significance of the conceptual framework, Vivian 386v and San Paolo 310v should be accorded primary status for source identification. In weighing their relative 58 On the whole, as far as the Utrecht Psalter is concerned, the building depicted in Utrecht 62v is the closest approximation of the Junius palace, though face-gabled in orientation, roof only on the top, and no definite indication of a clerestory.

82 

 3 The Fall of the Rebel Angels: P3

importance, one should address squarely the fact that they are largely in complementary distribution (7, 8, 9). Accordingly, they each must be acknowledged as complementary and mutually reinforcing sources. By contrast, Vivian 3v and San Paolo 3v are fully reducible to Vivian 386v and San Paolo 310v, respectively. Not a single feature is attested that owes solely to Vivian 3v or San Paolo 3v without attending involvement of Vivian 386v or San Paolo 310v, whereas the latter two are found in effect where the former two are absent. Therefore, Vivian 3v and San Paolo 3v would have been implicated secondarily, as their presumable participation is fully derivable from their inherent dependence on Vivian 386v and San Paolo 310v, respectively. As for Utrecht 26r, two of the three properties that Utrecht is presumed to have contributed to the illumination of Junius’s archetype are foreign to Vivian/San Paolo. In this light, the Psalter must be posited as an original, that is, nonderivative, model, along with Vivian 386v and San Paolo 310v. This does not mean, however, that Utrecht 26r must be characterized as equal to Vivian 386v and San Paolo 310v in breadth and depth of influence. Of revealing interest in determining the role of Utrecht 26r is the distribution pattern it evinces in distinction from Vivian/Paolo. Specifically, all of the three features that are demonstrably attributed to Utrecht 26r are the three major constituents of the center (Lucifer and the throne) and the right periphery (the upper and lower groups of angels). In other words, Utrecht 26r has no definite bearing on the composition of the left periphery. In this respect, Utrecht 26r diverges from Vivian 386v and San Paolo 310v. The latter two are no less involved in the left side than in the right. The presence of Vivian/San Paolo is thus ubiquitous throughout the register, whereas Utrecht is limited in its domain of influence. This radical difference may be subject to a principled explanation by postulating the divergent ways in which Vivian/San Paolo on the one hand and Utrecht on the other were brought to bear on the illustration of P3.1 (and the whole page for that matter; see below). Topographically, the Utrecht-originated material that was drawn on for production of P3.1 is confined to a compact terrain on the upper right zone ranging from hilltop to hillside. Furthermore, the features unique to Utrecht 26r in their distribution are readily accessible to a superficial observation—the gifts being carried and the throne placed outside the palace. Their identification is accomplished upon recognition of the individual objects involved, without exploring deeper meanings that may be decoded from them through their articulation with other figures present or absent, or without analyzing the items thus identified into smaller constituents and categorizing them in turn. In this light, it may be inferred with reasonable certitude that the influence of Utrecht 26r on P3.1 was mediated simply by a casual, shallow contact with the Psalter. The Junius artist would have been prompted to pick it up more by fancy than by careful deliberation and rational reasoning. Diametrically opposite is the situation involving Vivian 386v and San Paolo 310v. Extending beyond a strictly localized zone, the models’ effects are appreciated as permeating the entire panel (Table 7). On the one hand, by analysis the artist breaks into constituent pieces the level of individuated figures and entities that are most suscep-

3.2 A diachronic–comparative perspective 

 83

tible to gestalt identification. On the other hand, by abduction and generalization he reconstitutes a set of compositional and interpretive rules that seem to govern directly observable representations in the models. Thus, by adaptation and reorganization, the artist reuses the acquired working knowledge for his objective of representing a novel series of episodes that were far removed from the original. This innovative reuse of not only the fine details of surface representations but also, more critically the underlying conventions of organization would hardly have been viable without an intimate familiarity with and an analytical knowledge of the sources. Through close and intensive contact with the Touronian materials, the artist must have familiarized himself with the practice of the School of Tours in general as epitomized in the Vivian Bible and the Touronian preform of the San Paolo Bible. The artist’s access unlikely would have been limited to a particular single work accidentally at hand, such as the Vivian Bible. In view of the analogous register-based arrangement of pictures corresponding to narrative development, the identification of the primary sources of the first register of P3 as the initial registers of Vivian 3v/386v and San Paolo 3v/310v may lead one to expect comparable genealogical affinities between the remaining panels of P3 and those of the Touronian Bibles. As it turns out, the expectation is empirically borne out. Underlying P3.2a one may well recognize Vivian 3v3 and San Paolo 310v3 (left), in both of which the central figure, glancing leftward, is spreading his arms towards three men standing on both sides. More specifically, on the one hand, Lucifer receiving or delivering palms to the angels standing on both sides is matched with Jerome handing over manuscripts to the flanking monks in Vivian 3v3. There are two notable differences, however. Jerome is seated on the bench and only the two monks—rather than all six—in immediate proximity are facing the central figure. On the other hand, these two differences are lacking in San Paolo 310v3: Paul is standing, as are the six men who all look to him. Furthermore, the canopy over Paul is reminiscent of Lucifer’s wings in P3.2a. Instead of these closer similarities to P3.2a, however, there is an outstanding discrepancy in evidence: Paul is not engaged in the act of giving or receiving something tangible like palms or manuscripts; he is simply giving words, as he is involved in preaching. On the whole, then, P3.2a may be characterized as an adaptation of Vivian 3v3 and San Paolo 310v3, whereby the symmetrical arrangement of seven men, with the focal figure flanked by three on either side, is reproduced with variation through an eclectic selection of finer features that are in complementary distribution in the models. Of related significance, there are reasons to assume that both pictures—Vivian 3v3 and San Paolo 310v3—underwent recomposition before they were fixed in their present form. Both of the recompositions were presumably induced by the same motivation of composition, an inclination for symmetrical arrangement of figures. As for Vivian 3v3, Jerome’s simultaneous, symmetrical delivery of the Vulgate copies to the two monks on both sides, as depicted in the Vivian Bible, would have resulted through elaboration of the original simpler scene in which Jerome hands a copy to one monk, as a comparison with a cognate representation in San Paolo 3v3 (right) suggests, and more substantively

84 

 3 The Fall of the Rebel Angels: P3

as Kessler (1977: 86–87) argues drawing on a collection of presentation scenes at large like Hrabanus Maurus’ De laudibus sanctae crucis (Vienna, Österreichische Nationalbibliothek, Cod. 652), fol. 2v. As far as San Paolo 310v3 (left) is concerned, a closely affiliated representation—Vivian 386v3 (for details, see the treatment of P3.2b below)—and also other cognates like Epistulae Pauli (Munich, Bayerische Staatsbibliothek, Clm 14345), fol. 7v (Kessler 1977: fig. 182) and Rotulus (Vercelli, Bibliotheca Capitolare; Kessler 1977: fig. 184) indicate that originally Paul would have been shown addressing a single group of men standing on one side only (Kessler 1977: 114–116). Again, the elaboration aimed at a symmetrical composition is demonstrably in operation in San Paolo 310v3 as well. The above observations on the innovative symmetrical composition characteristic of the Touronian Bibles sheds new light on the relationship between Junius 11 (or more precisely its predecessor) and the Touronian Bibles. Since the presence of this feature in San Paolo is open to two competing accounts on its provenance (either it was inherited from the Touronian model or it emerged originally in San Paolo itself), it may be most appropriate to focus on Vivian out of caution. It may follow accordingly that the production of P3.2a in the archetype of Junius 11 postdates the Vivian Bible and hence exhibits fuller effects of symmetrical composition, an artistic practice that seems to have gained increasing currency in the School of Tours. At this point, Utrecht 89v (Pl. 82)—not 26r—comes into the picture (Rosenthal 1974: 188–189; Broderick 1978: 102–103; 1983: 166–167). The involvement in the same page of Junius 11 of a totally different picture (Utrecht 89v) that seems to have no intrinsic relation to Utrecht 26r seems to suggest that the Utrecht Psalter unlikely would have played such a formative role in the composition of P3 as did the Vivian and San Paolo Bibles. Particularly at stake is the bottom register of Utrecht 89v that visualizes a verse in “Gloria in excelsis Deo” (Lk  2:14), specifically, “in terra pax hominibus bonae voluntatis” (Peace on earth to men of good will; DeWald 1932: 70; Utrecht Psalter 2015: 209). At first glance, the resemblance to P3.2a exhibited by Utrecht 89v is outstanding indeed. The motif of the central angel grasping palms is commonly distinctive of P3.2a and Utrecht 89v, to the exclusion of Vivian 3v3 and San Paolo 310v3, thereby suggesting their privileged relation. In other respects, however, the two pictures diverge, rather than converge. Of further significance, on these discrepancies from Utrecht 89v, P3.2a agrees with Vivian 3v3 and/or San Paolo 310v3. First, in Utrecht 89v the two groups of men on both sides are too large to give individuality to each member and hence to figure out their numbers. As a visualization of the verse “Peace on earth to men of good will,” at issue is the collectivity of men of good will or the abundance of good men. By contrast, exactly three beings are represented on either side in P3.2a, Vivian 3v3, and San Paolo 310v3. Second, while the angel alone holds palms in Utrecht 89v, all the figures depicted in the panel carry palms or their equivalent (Vulgate copies).59 Third and related to the second point, whereas the two angels or men immediately at the left and right of the

59 No concrete objects are featured in San Paolo 310v3, as noted above.

3.2 A diachronic–comparative perspective 

 85

central one share the palms or codices with him in P3.2a and Vivian 3v3, the corresponding men do not partake of the items in Utrecht 89v, which are grasped exclusively by the middle-standing angel. Thus, it is only the palm-grasping angel at the center conceived in the abstract that securely relates Utrecht 89v to P3.2a. All remaining details of Utrecht 89v fail to find their way into P3.2a, even including the angel’s marked posture of grasping the palms crossed at the top. This may imply that the influence of Utrecht 89v remains superficial, limited as it is to the surface immediately accessible to a cursory observation. In this way, while the influence of Utrecht 89v on P3.2a is hardly in doubt, it cannot have been a well-motivated adoption and a well-designed adaptation. The Junius illuminator would have come across it more by chance than by careful reasoning. The core and bare image of the palm-carrying angel that may be ultimately derivable from Utrecht 89v was then materialized in the Touronian context, largely shaped by the practices that had been executed in Vivian 3v3. Accordingly, without full respect to the details of the central angel, its original remarkable way of grasping the palms was not replicated: it was replaced by a kinesthetically more natural posture of delivery. The recourse to Vivian 3v3/San Paolo 310v3 as the models of P3.2a explains why the central figure is flanked by three figures, not more and not less. Resituating the original images found in the contexts of delivering (Vivian 3v3) and preaching (San Paolo 310v3) would have motivated all noncentral figures to share the items being distributed in P.3.2a. Behind P3.2b, one may detect Vivian 386v3 as its primary model. In both, the central figure, God or Paul, attended by two followers, confronts three angels or men in preaching. In both, the central figure holds a scroll in his hand or arm. While basically corresponding to Vivian 386v3, San Paolo 310v3 exhibits some notable discrepancies. Paul stands between two groups of men, who are symmetrically distributed with three men placed on both sides. There is another dimension of symmetry worthy of note. That is, Paul addresses his preaching to both bands at the same time. Thus, the three men behind Paul are being no less challenged by the preacher than those in front. A still further symmetry concerns Paul’s posture of hands. Unlike Vivian 386v3 and its close cognate, P3.2b, Paul does not hold a scroll or any other item. Rather he raises both empty hands up toward the two groups in symmetrical fashion. Thus, the San Paolo illuminator’s attachment to symmetrical composition seems to have been so persistent that it induced a radical reconfiguration of the originally far simpler depiction that is more faithfully reflected in Vivian 386v3. While P3.2b is no doubt closer in affinity to Vivian 386v3 than to San Paolo 310v3, there are a few significant disparities between the former two. Most conspicuous among them, P3.2b and Vivian 386v3 constitute mirror images of each other. The central figure faces to the left (P3.2b) and right (Vivian 386v3). Correspondingly, the two cohorts stand to the right (P3.2b) and left (Vivian 386v3) of the main disputant, with the three antagonists positioned on the opposite side. Analogously, God holds a scroll in the left hand, whereas Paul does so in the right. This perfectly mirror-imaged configuration in P3.2b as opposed to Vivian 386v3 should be ascribed to the reversal in the direction of the central

86 

 3 The Fall of the Rebel Angels: P3

figure’s face (God’s in contrast to Paul’s), which in turn is derived from the reversal in narrative movement in P3.1 (right to left) in distinction from that in Vivian 3v1 (left to right), as discussed in detail above. Thus, the outstanding difference at issue proves to be an epiphenomenon, fully derivable from the basic orientation at work in the Junius pictures in P3 and beyond, namely, P2. It should be noted in passing that the apparent asymmetry between right and left in both P3.2b and Vivian 386v3 may be resolved by a symmetry at a higher level, that is, the one pertaining to all six figures: the three proponents (a leader and his two attendants together) confront three opponents. The remaining differences are minor and readily explainable. They may be adduced as confirming rather than contesting the genealogical affinity between P3.2b and Vivian 386v3 being proposed here. The palms carried by the angels correspond to the spears held by the soldiers. In light of the context immediately following P3.2a, the angels are expected to hold palms. The replaced spears, then, are transferred to God’s hand. Given the three antagonists to the left, God clutches three javelins in confrontation. A direct confrontation scene naturally leads to the posture of holding a weapon high and aiming it forward, rather than holding it steadily upright as in Vivian 386v3. Consequently, whereas Paul holds nothing in his left hand, God holds up the three javelins in his right hand instead. Moreover, the jointed border, which may be conceptualized as the ceiling as observed in the preceding section, may well be regarded as vestiges of the model scene depicted in Vivian 386v3, in which Paul’s confrontation takes place under the roof, rather than outside. The event is depicted in an indoor setting also in the corresponding picture of San Paolo 310v3 (left). In this way, the foundational composition of P3.2b is mapped from Vivian 386v3, while all the attending minor discrepancies between the two pictures are subject to explanation as derivatives of the well-motivated reconfigurations executed in the process of mapping. Alternatively, one may be led to view Utrecht 37v (Pl. 76; Raw 1953: 71) or Utrecht 41v (Pl. 78; Broderick 1978: 103–105; 1983: 167) as most probable sources of P3.2b. The relevant picture in Utrecht 37v is described as follows: At the top of the picture the beardless Christ-Logos without nimbus is riding out of the clouds in a quadriga suggested by the ‘chariot’ of verse 18, 17, flanked by four angels, three of them bearing wands. The Christ-Logos brandishes a torch and is ‘scattering’ His enemies before Him who “flee” from His presence. (verses 2–3, 1–2; DeWald 1932: 31; see also Utrecht Psalter 2015: 171)

This drawing is a composite visualization of Psalm 68:1, “Let God rise up, let his enemies be scattered; let those who hate him flee before him” (Coogan 2010), and Psalm 68:17, “With mighty chariotry” (Coogan 2010). A more striking analog of P3.2b is Utrecht 41v in Broderick’s view. “[A] wingless angel holding a whip and [three] spears (the whip suggested by ‘flagellabuntur’ [‘scourged’] and ‘flagellatus’ of verses 5 and 14), is driving a large number of the wicked into a fiery pit of Hell in which appears the flaming head of Death (verse 18)” (DeWald 1932: 33; see also Utrecht Psalter 2015: 173). This is a visual representation of Psalm 73:27: “Indeed, those who are far from you will perish; you put an end to those who are false to you”

3.2 A diachronic–comparative perspective 

 87

(Coogan 2010). Furthermore, six other angels are represented in the sky above the angel to the right, with groups of three standing face to face. Apparently unrelated to any particular verse, however, these six angels hardly may be regarded as being in company with the javelin-bearing one. Whichever of the above two Utrecht representations might resemble P3.2b more closely, it is worth emphasizing that the similarity involved does not solely concern P3.2b; rather, it extends to the depiction of hell as represented in P3.3. In view of such differing domains for comparison, the choice between Vivian/San Paolo on the one hand and Utrecht on the other as sources of P3.2b should not be necessarily a matter of mutual exclusiveness. With this qualification in mind, it is appropriate at this juncture to focus on evaluating the Utrecht pictures against each other in strict relation to P3.2b. Utrecht 37v and 41v are largely complementary in their resemblance to P3.2b, as indicated in Table 8. In 37v, God, riding on a chariot and holding a javelin, is attended by two angels on either side. In 41v, an angel standing alone holds up three javelins, while far above in the remote background three angels stand on each side facing each other. Given the neat complementary distribution of variant features in the two pictures, one may be led to surmise that the artist would have drawn on both pictures in an eclectic manner, ending up eventually with the composition of P3.2b. Table 8. Distribution of P3.2b features in Utrecht 37v and 41v  

P3.2b

Utrecht 37v

Utrecht 41v

God 2 angels at right 3 angels at left Attending angels Standing on feet Three javelins Scroll

√  √  √  √  √  √  √ 

√  √    √       

    √    √  √   

This view on the hybrid sources of P3.2b, however, leaves a few significant properties of the Junius scene unaddressed. First, the arrangement of three and two angels at left and right, respectively, remains to be explained. All that can be accommodated is the asymmetrical distribution of two and three, not exactly the actual arrangement achieved— two rather than three chosen as God’s followers (right) and three rather than two as God’s opponents (left). Moreover, it remains unclear why asymmetry would have come to prevail in defiance of the original symmetry. Second, and most importantly, there is a striking discrepancy between P3.2b and Utrecht 37v/41v. At issue is the direction in which the central figure faces, as is epitomized by his line of sight. While the main character looks downward at the falling enemies in either picture of the Utrecht Psalter, the corresponding figure in P3.2b seems to be gazing at the group of angels on the left,

88 

 3 The Fall of the Rebel Angels: P3

particularly the one standing in front. In other words, God or the chief angel primarily connects with his enemies below in the Utrecht representations, whereas he is principally linked to the angels standing on the same ground. That is, the main force of combination works vertically in Utrecht 37v/41v, but horizontally in P3.2b. The primacy of horizontal link in the latter representation is reinforced by the thick border between P3.2b and P3.3. By contrast, nothing comparably conspicuous separates the two layers in the alleged Utrecht models. In order to provide a principled account of these unique properties of P3.2b, one must invoke Vivian 386v3 as an explanans. As may be recalled, Paul, with two supporters standing behind him, is engaged in dispute with the group of three men to the right. The precise number and arrangement of participants in the preaching, as well as the confrontation itself, may explain most naturally the specific composition realized in P3.2b. Therefore, one may be justified in concluding that whatever role the two Utrecht pictures played in the composition of P3.2b would ultimately have been governed in their activation by the creative force that had shaped Vivian 386v3.60 The critical involvement of Vivian 3v/386v and San Paolo 3v/310v in the composition of not only P3.1 but also P3.2a and P3.2b may well lead to the prediction that they would have been no less crucially implicated in the creation of P3.3. The prediction seems to be borne out at first sight. Most prominently featured in P3.3 are obviously Lucifer falling and Satan lying and bound. These two focal figures are formally derivable from the falling Saul and the sleeping Ananias, respectively, in Vivian 386v1/2 and San Paolo 310v1/2. It may be worth noting further that the two analogously postured figures—Ananias and Satan—are lying at the bottom of the walled enclosure in Vivian 386v2 and P16.2, a variant of P3.3 (section 3.1; see also below). While the two distinct individuals—Saul and Ananias—are involved in the alleged models in contrast to the single identity—Satan—of their Junius descendants, the two Touronian antecedents are intimately associated with each other specifically by virtue of the common presence above them of God’s hand and the ray of light emitted from it, apart from the analogous postures they share. The association of falling and lying figures, however, is no less viable in Utrecht 37v/41v than in Vivian/San Paolo. As far as Utrecht 37v is concerned, several men are represented lying in sepulchers farther below, corresponding to Psalm 68:6 (67:7), “but the rebellious live in a parched land” (Coogan 2010), although this visualization is originally unrelated in content to the fallen enemies. As for Utrecht 41v, three men are depicted lying in bed, a representation of Psalm 73:20 (72:20), “They are like a dream when one awakes” (Coogan 2010). Again, the awakened men have conceptually nothing to do with God’s opponents falling, depicted immediately below them. The contiguity

60 As will be discussed in the following, the contributions that the Utrecht 37v/41v made to the representation of P3.3 are indisputable. In this light, it would be inadmissible to dismiss the Utrecht pictures entirely from the formation of P3.2b and identify the Vivian counterpart as the sole model at work.

3.2 A diachronic–comparative perspective 

 89

of the two distinct images, however, would have been instrumental in associating and reconfiguring them as a unified whole in the two pictures of Utrecht. While the derivability of P3.3 from Utrecht 37v/41v thus detracts from the plausibility of identifying Vivian 386v/San Paolo 310v as the principal sources of P3.3 in regard to the falling and lying Lucifer, P3.3’s overall similarity in composition to the lower part of Utrecht 37v and 41v—particularly the latter’s depiction of the falling crowd—undermines the wholesale reducibility of P3.3 to Vivian 386v and San Paolo 310v at the expense of the Utrecht representations. The two Vivian/San Paolo pictures contain no images remotely resembling the falling mass of people represented in the Utrecht counterparts. Since there seem to be no specific elements that correspond exclusively to Vivian/San Paolo and accordingly defy reduction to the Utrecht pictures, one might be tempted to conclude that Vivian/San Paolo’s unique contributions to P3.3 cannot be demonstrably sustained; therefore, the sources of P3.3 should more convincingly be determined as Utrecht 37v/41v. On closer inspection, however, things come to be seen in a different light. At stake is the walled enclosure of hell, the building structure that may be viewed as analogous to Ananias’s house and the City of Damascus in Vivian 386v2 and San Paolo 310v2. Absent in Utrecht 37v/41v, this enclosure is missing also in P3.3, to be sure, but is realized in the two variant representations of P3.3, namely, P16.2 (Pl. 10) and P17.2 (Pl. 11). The absence in P3.3 on the one hand and the presence in the latter two on the other of the walled representation of hell are motivated by the differing perspectives involved. As remarked in the previous section, the three representations at issue concern three different temporal phases of the event of the Fall of the Rebel Angels in such a way that the three scenes may be sequentially ordered: P3.3 > P16.2 > P.17.2. Thus, P3.3 focuses on the process of falling in itself through the air to the mouth of hell. P16.2 broadens the perspective to include the situation that obtains immediately after passing through the hell mouth. Lastly, P17.2 is concerned with the final phase of the event, a perpetuated torment in hell, in contrast to the orderly presence of God in heaven (P17.1). In this light, inasmuch as the structure of hell falls outside the scope of interest in P3.3, the walled hell does not figure at all in the representation. This would not mean, however, that it is nonexistent; rather its existence is being implied below the mouth, as it is manifestly depicted in the other variant scenes. The immanent presence of the unrepresented walled hell in P3.3 may legitimate an exploration of its sources and derivations on the basis of its manifest representations in P16.2 and P17.2, and warrant a postulation of the results as part of inherent properties of P3.3 by extrapolation. The walled hell in P16.2 and P17.2, it must be emphatically pointed out, is similar in shape and composition to the walled towns in Vivian 386v2 (right) and San Paolo 310v2 (middle). Specifically, the overall hexagonal shape, as most clearly illustrated in Vivian 386v2, seems to be still appreciable in the Junius pictures in question. Some of the towers attached to the walls—the extreme right and left ones— are obviously reconfigured as pillars in P16.2, but still retained as such in P17.2. The wall between towers on the two corners farthest from the viewer would have been redesigned in P16.2 as the hell mouth, the jaws being vestiges of the two towers. Of

90 

 3 The Fall of the Rebel Angels: P3

special interest are the wall bricks with a single circle at the center, which are commonly depicted in P16.2 and Vivian 386v2 to the exclusion of the San Paolo picture. The two board-like entities projecting from the walls on the viewer’s side in P16.2 are reminiscent of the doors located in the corresponding places in Vivian 386v (right and left). While the door is closed in Vivian 386v (right), it is half-open in the left picture. Moreover, this open door, while not of the town wall, is situated at the feet of Ananias sleeping, a configuration that is analogous to Satan lying with the feet placed closest to the open gate in P16.2. It may be inferred then that the enigmatic items in question are vestiges of the open front doors, which would have resulted through a mutual assimilation of the two doors represented in Vivian 386v2 at right and left. Of no less importance than the walled structure and its fine architectural attributes is the depiction of Satan and his band of fallen angels enclosed in hell in P16.2 and P17.2. Focusing on a pair of facing figures in P17.2, namely, Satan at right and one of the angels confronting him at left, will bring into relief a parallelism in composition to Vivian 386v2 (right), in which Saul at right and Ananias at left are faced with each other. Further agreements in detail are noteworthy. The right figure (Saul or Satan) is positioned relatively lower than the left one (Ananias or the angel); Saul or Satan is looking up to his companion, who holds out his hand over Saul’s or Satan’s head. P16.2 provides a variation on the configuration in P17.2: parallel to Ananias in Vivian 386v2 (left), Satan is lying, rather than sitting up as in P17.2. Despite Satan’s deviant posture, P16.2 exhibits a significant point of agreement with Vivian 386v2, which is unshared in P17.2. That is, the angel confronting Satan stands on the front-left tower. Indeed, Ananias does not assume the same posture in Vivian 386v2, but he is positioned just beyond the corresponding tower. The proximity of the angel/Ananias to the left tower this side, then, may count as another common feature that affiliates Vivian 386v with the triad of P3.3/ P16.2/P17.2.61 The focused presence of Satan among the crowd of falling and fallen angels in all three variant Junius pictures may in turn cast doubt on their wholesale derivation from the Utrecht sources. In Utrecht 37v/41v, no figure can be identified as a central character among the doomed group. The privileged representation of Satan accordingly would be amenable to explanation by hypothesizing that underlying the triad pictures are not only the Utrecht models, but also Vivian 386v2/San Paolo 310v2 at a deeper level. The latter would have served as the foundational inspiration, whereas the Utrecht Psalter would have provided auxiliary material for compositional elaboration. The above argument might provoke the counterargument, as follows: motivated in textual terms, the privileged visualization of Satan is something that is secondarily added to the background of the falling foes, the foundation that was in itself inspired 61 San Paolo 310v2 (center) constitutes a mirror image of Vivian 386v2, whereby Ananias and Satan are reversed right and left. This is a consequence of placing the scene of the sleeping Ananias on the right side of the register, at the expense of narrative development and in favor of symmetrical composition (cf. Kessler 1977: 114).

3.2 A diachronic–comparative perspective 

 91

exclusively by Utrecht 37v/41v. Thereupon, one is brought back to the initial observation one started with: in what seems to be the original image to be reconstructed as underlying the triad of P3.3/P16.2/P17.2, Satan is represented seated on the right side, confronted by another angel to the left who is positioned closest to the left tower and relatively higher than Satan, and who holds up his hand over Satan’s head. This configuration turns out to be isomorphic in its essentials to Vivian 386v2 (right), in addition to being enriched with a variation corresponding to Vivian 386v2 (left). If the central figure had been specified derivatively as a contingency according to the counterargument in question, there would be no reason why it should exhibit significant agreement in composition with what is depicted in Vivian 386v2. Inasmuch as one adheres to Utrecht 37v/41v as the principal sources of P3.3/P16.2/P17.2, one is bound to fall short of accounting for the otherwise explainable correspondences in composition. In terms of iconography and composition, Vivian 3v, Vivian 386v, San Paolo 3v, and San Paolo 310v contributed to the construction of P3 as a whole in their differing and complementary ways. Of these sources, Vivian 386v was by far the most significant. The three registers of Vivian 386v would have been mapped onto P3 through reconfiguration and rearrangement, whereby the City of Jerusalem (Vivian 386v1), along with the City of Rome (Vivian 3v1), Jerome’s Hebrew teacher’s school (Vivian 3v1), and Ananias’s house (Vivian 386v2), are transformed to Lucifer’s palace (P3.1); Saul’s astonished guards (Vivian 386v1 and San Paolo 310v1) and Jerome beckoning (Vivian 3v1 and San Paolo 3v1), to Lucifer inviting (P3.1); the fall of Saul (Vivian 386v1 and San Paolo 310v1), to that of Lucifer (P3.3, upper); Paul’s disputation (Vivian 386v3), to God’s attack on the rebel angels (P3.2b); the healing of Saul in the enclosed place and the sleeping Ananias (Vivian 386v2) in his house, to the bound Satan in the walled hell (P16.2 and P17.2). In addition, Jerome’s distribution of Bible copies in Vivian 3v3 would have served as a primary model for P3.2a. In regard to the overall organization of P3 for illustration, the strict register-based framework and the arrangement of a single scene per (sub)register is susceptible to explanation on the basis of the Vivian model alone, especially Vivian 3v, in which the original multiplicity of scenes in a single register is minimized in favor of a single scene representation (Vivian 3v2 and 3v3). By contrast, while striking on the surface, the resemblance to the Utrecht Psalter is no more than a matter of superficial similarities. In brief, Utrecht serves as a repository of raw materials for reuse, 26r for P3.1 and 37v/41v for P3.3. It may be worth pointing out further that the influence of Utrecht concerns depictions of large assemblages of figures. The exact ways of reuse and reorganization, however, are determined primarily by the conventional practices characteristic of the Touronian Bibles.

4 The Creation: P6/P7 4.1 A synchronic–structural perspective P6 (Pl. 4) and P7 (Pl. 5) constitute a tripartite visualization of creation, each of the three constituent pictures devoted to two days of creation. The composition of the three pictures is identical: they consist of two semicircles, the upper half standing for the heavenly sphere, and the lower dome the terrestrial one. P6 depicts the separation of light from darkness on the first day and the raising of the firmament on the second day. Rays of light are pouring from a vessel carried by an angel who is hovering at the top of the outer arc. The Creator is seated on top of the inner semicircle, which represents the firmament. Down below and on the surface of deep water—the domain of darkness—is seen the upper body of another angel, who covers its face with veiled hands as if to embody literally “tenebrae erant super faciem abyssi” (darkness was upon the face of the deep; Gen. 1:2; Raw 1976: 136) and/or at the same time to avoid dazzling light emitted from above, thereby epitomizing the overwhelming effects of light on darkness. The newly created light thus encroaches on the darkness that was of primeval omnipresence. The two angels may accordingly be regarded as personifications of light and darkness, as interpreted by Raw (1955) and followed by Blum (1976: 224n27) and Broderick (1978: 112–117, 140). Giving justice to the chronological order of creation, P7.2, the lower picture on page 7, is concerned with the third and fourth days of creation. As illustrated in the lower semicircle, plants grow on earth (the third day), and then the stars are set and distributed over the sky (the fourth day). Noteworthy are two related events that failed to be represented. On the one hand, the separation of land and water is not explicitly depicted, but simply presupposed in the immediately following phase, the creation of plants. On the other hand, the sun and the moon—the major and minor great lights— are not specifically illustrated. In the upper dome of P7.2, the Creator is shown standing, and overhead an angel holds a bowl–like source of light, suspended from the top of the dome, basically the same motif as in P6. Finally, P7.1 the upper half of P7, visualizes the creation of birds (the fifth day) and animals (the sixth day). Oddly enough, fishes and other sea creatures, which were brought into being on the fifth day of creation along with fowls, are not given a specific depiction, although the waters are apparently illustrated below the land, thereby possibly inviting inference as to the creation of fishes. Above the lower dome is God shown standing, essentially the same posture as in P7.2 but enclosed in a mandorla and unaccompanied by a bowl–holding angel. Attached to the P6 and P7 pictures on the margin are two inscriptions, one on each page. The first one is found in the margin to the left of the top of the upper arc in P6: “[Hēr hē] g[e]syndrode wæter and eorðan” (Here he divided the water and the earth; Gollancz 1927: xxxix). The first five letters are supplied by conjecture, largely on analogy to the inscription on the following page, since whatever would have preceded the first https://doi.org/10.1515/9783110788068-004

94 

 4 The Creation: P6/P7

legible letter was obliterated as a consequence of binding. The second inscription is provided on page 7 to the right of the second disc: “Hēr hē tōdǣlde dæg wið nihte” (Here he divided the day from the night; Gollancz 1927: xl). Standing largely mirror-imaged, these inscriptions are placed in symmetrical fashion on the facing pages. In general terms, these pictures seem to be readily amenable to a straightforward account, as outlined in the preceding paragraphs. In specific details, however, there are notable points of disagreement. Most contestable are the double occurrence of plants, and their distinct manners of representation, in P7.1 and P7.2. Not only plants are depicted on the two illustrations (as if to misleadingly suggest that vegetation is produced twice in the course of creation), but they are represented in conspicuously distinct form: in impressionistic terms, the plants look flourishing and exuberant in size and shape in P7.1, in contrast to the P7.2 counterparts, which appear rather bleak and feeble. According to the reading presented above, P7.2 covers two days of creation, much as the other two pictures do: the creation of plants (the third day) and the setting of stars on the dome of the sky (the fourth day). Therefore, the representation of vegetation in P7.1 is an inheritance from an earlier day (the third day) when it was brought into being. The apparent lack of animation in the plants represented in P7.2 may be ascribed to their incipiency in existence and/or to their minimal activity at night. Plausible as it is, this interpretation, which was earlier presented by Blum (1976) and Broderick (1978), and on which I am now elaborating, is formulated as a better alternative to the more common view that still has wide currency through authoritative works such as Gollancz (1927) and Ohlgren (1992): the creation of plants is represented in P7.1, along with that of birds and animals. In this view, P7.1 is concerned with the third day as much as the fifth and sixth (Gollancz 1927: xl; Ohlgren 1992: 89; Finnegan 1998: 23). Correspondingly, P7.2 is devoted exclusively to the fourth day of creation, the separation of day and night. Gollancz’s and others’ reading, however, is obviously at variance with the appearance of vegetation in P7.2, much less capable of accounting for their form being markedly distinct from that in P7.1. As observed above, the fragile appearance may well suggest that the plants in P7.2 were designed to denote a younger stage than those in P7.1. That is, the plants in P7.2 are closer to their emergence on land. Accordingly, their less developed appearance should be more naturally associated with the third day of creation than their fully matured forms implying a more advanced stage than at their original creation. Furthermore, the standard interpretation is incapable of explaining the unnatural alignment that it makes between the three representation spaces of P6 and P7 on the one hand and the six days of creation to be represented on the other. At stake are the following two peculiarities. First, given that three registers of approximately the same size are available for depicting the six days of creation as a whole, it would seem most natural that each zone be allocated two days for representation, that is, the correspondence of one register to two days, as assumed by the interpretation advocated by Blum and Broderick as an alternative to the common conceptualization. The standard interpretation, by contrast, claims that the six days are unevenly distributed in such a way that the first/

4.1 A synchronic–structural perspective 

 95

second days (P6), the third/fifth/sixth days (P7.1), and the fourth day (P7.2) are treated in the three distinct spaces. Second, given that the six days of creation are strictly ordered in temporal succession from the first to the sixth, the three depictions are expected to be arranged in a correspondingly continuous progression even if the optimal alignment of two days with one register is not implemented. In this respect, too, the common view must postulate a deviation from the otherwise expected conformity with the strict temporal sequence, whereby nonsuccessive days—the third, fifth, and sixth days, are put together as a unified scene, to the exclusion of the fourth. As a consequence, P7.1 and P7.2 exhibit a contradiction with the chronological order by placing the scene of the fourth day after those of the fifth and sixth. In other words, the conventional view would have to impose a complicated back and forth movement between P7.1 and P7.2 on the part of readers/viewers. In modifying the above handbook account, Raw (1976: 136–137) assumes that P6 represents exclusively the first day. On this basis, she generalizes that P6 and P7.2 are both concerned with the separation of light/day from darkness/night. Raw thus privileges the opposition between light and darkness as a foundational theme and allocates it its own representation space on both sides among the tripartite zones available, namely, P6 and P7.2. More specifically, according to Raw’s conceptualization (1976: 136), P6 and P7.2 constitute outer panels, between which P7.1 stands as the middle panel. Such a tripartite construct, she further assumes, originates from an ivory triptych. In highlighting further a parallelism between P6 and P7.2, Raw (1955: 318) brings to light a compositional symmetry: much as in P6, the bowl held by the angel—the source and symbol of the light/day, implying the sun—is represented in the upper arc in P7.2, whereas the darkness/night, symbolized by the stars, is allocated to the lower dome.62 Consequently, all the remaining days of creation—except for the second day, about which, see below— are lumped together in the central space, P7.1, hence its heterogeneous nature in terms of objects represented. Raw’s revision may thus apparently provide an explanatory account of the two peculiar properties of P6/P7 mentioned in the previous paragraph, that is, the asymmetry in alignment and the discontinuity in progression. This explanation, however, is obtained at a great price: the treatment of the second day needs to be fully integrated with the new framework, as Raw left this issue utterly untouched. With varying degrees of persuasiveness, three possible solutions seem to be available, depending on whether the second day is actually featured in the tripartite creation pictures at issue, and if in the affirmative, where it is located, in P6 or P7.1. First, one may continue to assume, by compromise with the standard view, that P6 represents the first two days. The acceptance of this traditional view, however, undermines the very basis for Raw’s privileging of the light/darkness motif, since it no longer

62 This interpretation entails a disputable consequence of identifying the angel’s bowl as the sun. For a criticism of this point, see below in this section and the following one.

96 

 4 The Creation: P6/P7

exclusively occupies P6 for representation. Thus, it would hardly be viable to fully harmonize Raw’s thesis with the standard interpretation. Alternatively, one might assign the second day, along with all the remainder, to P7.1, except for the privileged pair of the first and fourth. At first glance, this interpretation appears to be plausible, in terms of conceptual consistency and economy. Empirically, however, it seems difficult to recognize in P7.1 a scene resembling the raising of the firmament on the second day, although the complete absence of celestial bodies on the lower arc in P7.1 might possibly be construed as standing for the sky in its emergence. Since Raw (1976: 136) regards the overturned bowl held by the angel as the source of light, its absence in the middle panel (P7.1) in contrast to its presence in the flanking ones (P6 and P7.2) may also be viewed as a reflection of the emergent vault of the sky before the positioning of the sun, the moon, and the stars on the fourth day. Such an interpretation, contrived out of absence of pictorial substance in support, is bound to remain pure speculation. A third solution would seem closest to what Raw would have articulated, if obliged to: the second day just failed to be treated in P6/P7. Raw’s rationale for this claim may be expressed in her own words: It is important to note that the creation pictures in Junius 11 are not of a narrative kind: they include most of [emphasis mine] the material from Genesis  1.1–25, but the events of the third, fifth and sixth days, with the exception of the creation of man, are grouped together in the central picture and framed by the two representations of the separation of light from darkness, which belong to the first and fourth days. (Raw 1976: 137)

As it turns out, the second day is exactly the exceptional part that eluded visualization in P6/P7. The characterization stated immediately above lacks explanatory power: it is nothing more than a stipulation, a quasi generalization derived from the preconceived absence of the second day from the creation pictures, which in turn is determined by Raw’s exclusive association of P6 with the first day in defiance of the presumable involvement of the second day as substantiated below. In this way, the exclusion of the second day from P6/P7 is a necessary consequence of Raw’s imposition of a particular interpretative framework on these pictures, rather than a conclusion based on firm empirical evidence. Moreover, granted that only the second day is missing in the creation pictures, as Raw would have it, it remains inexplicable why this day, rather than any other, escaped pictorial representation, despite its substantial treatment in the text (Genesis A 143b–153b). Notwithstanding Raw’s tacit assumption to the contrary, there is reason to believe that P6 represents the second day of creation as well as the first. Of vital importance in this regard is God’s positioning relative to the firmament. In P6, the Creator is seated on the viewer’s side of the top of the firmament, with the tips of his toes depicted in full. In contrast, in either picture of P7, the Creator is shown standing beyond the circumference of the sky, with his feet accordingly hidden behind the vault. This contrastive positioning may be construed as indicative of God’s differing engagement in creation.

4.1 A synchronic–structural perspective 

 97

The inclusion of the Creator in the inner, foregrounded arc in P6 in opposition to his positioning in the upper, backgrounded arc may well constitute an iconic representation of his total commitment to the construction of the vault of the sky, the inner dome, in accordance with the textual description. Denying the second day of creation its corresponding visualization in P6, as does Raw, would be challenged to account for why the opposition at issue in the Creator’s positioning obtains the way it does, rather than other conceivable patterns of distinction or an absolute lack of differentiation at all. The distinction between God sitting or standing is also subject to a comparable account. Inasmuch as a sitting posture incurs contact in a more extended area of the body, it is more appropriate to reserve this attitude for expressing a specifically intimate relation of the two parties involved, the Creator and the sky being created. Also explainable along a similar line of reasoning are the pointing–down gesture of God’s left hand, and the consequence of his not holding the codex in P6, as opposed to P7.1 and P7.2. While interpretable in more than one way in details, it seems legitimately conceivable that God’s pointing down to the surface of the waters serves to differentiate from earth the place where he is now being situated, up in the sky. Given that this lofty position was not in existence before God’s act of creation (Genesis A 103a–104a “Ne wæs hēr þā giet / nymþe heolstersceado // wiht geworden” [Then was nothing yet here except darkness; Anlezark 2011: 9]), the configuration of the Creator and the heavens, and the specific posture that he takes, may testify that the celestial expanse over the waters is of God’s creation. Therefore, it may follow as a matter of course that the second day of creation is represented assuredly in P6. Furthermore, the separation of the codex from God’s hand that is otherwise inalienable and its presumably consequential relocation under his feet represented in red (cf. Broderick 1978: 109) may seem to be compatible with the interpretation being proposed here that God’s special act, the act of creating the sky, necessitates a temporary unholding of the codex on this occasion. One may go so far as to conjecture that the resultant intermediacy of the scripture between the Creator and the firmament visualizes the phrase þurh his āgen word ‘through the intermediary of his own word’ in the text, Genesis A 147b–150a “and geworhte þā // roderas fæsten. / þæt se rīca āhōf // up from eorðan / þurh his āgen word, // frēa ælmihtig” (and then made the heavens, the firmament, which the powerful one, the Lord almighty, lifted up from the earth by his own word; Anlezark’s [2011: 13] translation). According to Henderson (1975: 140–143), the trio pictures of P6/P7 are united by the common subject of the opposition between light and darkness. P6 visualizes the separation of light from darkness on the first day of creation. Subsequently, P7.1 and P7.2, complementary to each other, represent the day and the night as embodiments of light and darkness, respectively, as they are brought into being on the fourth day of creation. On the whole, Henderson’s reading draws heavily on the text that runs before and after the P6/P7 pictures, in which, he claims, the division of light from darkness and their naming day and night are at issue, exactly the subject treated in P6 and P7, in Henderson’s (1975: 141) view. What is more, this interpretation turns out to be in harmony with the inscription on page 7 referred to above: “hēr hē tōdǣlde dæg wið nihte” (here he

98 

 4 The Creation: P6/P7

divided day from night). In light of Henderson’s account, this inscription constitutes an apt legend for the alleged theme underlying the two pictures (cf. Henderson 1975: 142). Despite this optimal text–picture correspondence, however, the first inscription on page 6, “[hēr hē] g[e]syndrode wæter and eorðan” (here he divided water and earth) must be acknowledged as an error (Henderson 1975: 141; see also Lucas 1980: 209). Thus, an appeal to the inscriptions to justify Henderson’s readings of P6/P7 has to remain inconclusive. Being a no less serious anomaly, the depiction of birds and animals in P7.1 would contradict the illustrator’s intention, as Henderson sees it, and hence would throw doubt on Henderson’s grand interpretation itself (Broderick 1978: 125): as of the fourth day—the day on which the day and the night come into being—these living creatures have yet to be produced. While Henderson is aware of this point of inconsistency in illustration, he minimizes its gravity by labeling it as “a mild and justifiable anachronism” (Henderson 1975: 143). Henderson’s view furthermore implies that the fifth and sixth days would properly have been illustrated on lost leaves. By contrast, Gollancz’s and others’ interpretations— whereby the P7 drawing constitutes a full illustration of the third through sixth days of creation—would entail that the three lost leaves would unlikely have contained separate pictures specifically designed for these later days of creation, as they are already dealt with in the existing pictures. In the absence of relevant data, these differing implications cannot be verified in one way or another. As the foregoing discussion has substantiated, the Blum–Broderick–Suzuki interpretation of P6/P7—presented at the outset of this section and elaborated on through comparison with competing accounts—is equipped with greater explanatory power than the Gollancz–Ohlgren–Raw reading, on the one hand, and Henderson’s on the other. Some residual issues have yet to be addressed, which appear to be challenging to the account being defended here. Some entities are treated apparently without existential consistency in P6/P7. First, the water is not depicted in P7.2, while it is properly represented in the other two pictures. Is this due to lack of space for further representation at the bottom of page 7? This can hardly have been the case in light of P6, which would suggest that P7.2 has space available for illustrating the water (albeit on a reduced scale). More plausible would be to make reference to the functional distinction between foreground (what is being focused on as significant information) and background (what is presupposed as being simply for presenting the foregrounded material). Since P7.2 is concerned with creation of plants on land, and since water counts as the binary opposite of land, it is relegated to the background in terms of information structure, and accordingly is not featured in illustration. By the same token, the sky in P7.1 is left blank without the stars, which, created earlier on the fourth day, as illustrated in P7.2, must be in existence out there on subsequent days. One may wonder if the open sky is intended to represent the day, as assumed by Henderson 1975. One is confronted with a further question: Why is the sun not featured at all, then? The complete absence of celestial bodies in general in P7.1 should rather be ascribed to a representation of the sky in a generalized, abstract form in which

4.1 A synchronic–structural perspective 

 99

the distinction between day and night is neutralized: the sky illustrated in P7.1 is neither of the day nor the night; it is the sky conceptualized in generic terms as a vast space. Such a neutral representation is in turn motivated by a marginalized status of the sky in the context of P7.1. The primary concern is obviously the creation of birds and animals on land, events that are conceptualized as independent of the separation between the day and the night. Accordingly, treated as a background object, the sky is defocused in visualization and therefore it receives minimal representation in its neutral, abstract form—the void firmament. In this regard, the treatment of the sun deserves further consideration. While Gollancz (1927: xl) and I assume that the sun is not represented at all for differing reasons (see above), Raw (1955: 318) and Broderick (1978: 124) are committed to a contrary view: the vessel of light held by the angel in P7.2 represents the day, hence by implication, the sun. This interpretation would mean that the sun should be located far above the firmament, in which the stars, the sun, and the moon are set on the fourth day (Gen. 1: 14–19), corresponding in part to the scene illustrated in P7.2. Such a spatial separation of the sun and other celestial bodies seems implausible, as they are all placed in the same sphere in other representations in the Cotton Genesis tradition, such as the San Marco mosaics (Demus 1984: pl. 113; Büchsel/Kessler/Müller 2014: pl. 8) and the Salerno Ivories (Pl. 53). More substantively, given that the lower arcs in P6 and P7.1 are reserved exclusively for representation of the earth and the sky, one would legitimately expect the same consistent arrangement for P7.2 in the absence of overwhelming evidence to the contrary. In these two pictures, the upper dome depicts God’s sphere, heaven, as sharply distinguished from this world, earth. Accordingly, contrary to Raw’s and Broderick’s conceptualization, it seems unlikely that the upper dome was meant to represent day (and by implication the sun), a phase characteristic of this world. Yet, seen from a diachronic–comparative perspective (section 4.2), it turns out to be more likely that the Junius artist deliberately treated the angel carrying light as ambivalent in representation, light in general on the one hand and the sun on the other: the two entities would appear to have been inseparably conceptualized in his mind. In anticipation of the argument in the next section, one may say at this point that while the angel carrying the light should not be regarded as directly standing for the sun, the symbol may be capable of referring to it by virtue of its encompassing domain of signification. Broderick (1978: 124n38) draws attention to an inept execution of the angel and the vessel in P7.2 compared with its P6 counterpart. Unless this is entirely attributed to the relatively smaller space available for the angel in P7.2 on account of the standing figure of the Creator, the observed difference may be understood as a reflection of the varying status and presence of light. Specifically, the bowl in P6 embodies the creation of light in separation from darkness. In contrast, the vessel in P7.2, inherited from the first day of creation as recorded in P6, may be reconceptualized as another of God’s attributes, heavenly glory, or “a sort of variation on the mandorla,” as Raw (1955: 318) aptly characterizes it. It should be interesting to recall that the standing figure of God is surrounded by a mandorla in the following picture (P7.1), implying the conceptual affinity between

100 

 4 The Creation: P6/P7

the two images. The difference in existential status between the two representations of light (P6 and P7.2) may be regarded as empirically confirmed by a greater range of rays of light emitting from the bowl in P6 than P7.2, in which the rays stream exclusively over the Creator. Thus, progressing from P6 to P7.2 to P7.1, the light in the upper arc shifts its reference—from light in the universe to light as an attribute of God—without a radical change in its corresponding image (P6 to P7.2), and then receives a divergent illustration without a comparable conceptual change (P7.2 to P7.1). The privileging of foreground entities in visualization, at the expense of those in the background, may thus provide an explanatory account of why some objects fail to be represented in the illustrations P7.1 and P7.2. Considerations of representational economy tend to curtail details from objects relegated to the background. This leaves a single puzzle about P7.1 still to be solved: fishes do not occur at all, although they belong to the class of foregrounded objects, namely, the beings created on the fifth and sixth days. Of relevance here is the faint rendition of what appears to be the water below the land in P7.1 (Gollancz 1927; Broderick 1978). Executed apparently with hesitation, this representation might well be regarded somehow as a reflection of the conceptually peripheral status that the illustrator accorded to the water as opposed to the land. Correspondingly, by way of speculation the failure of fish to be depicted in P7.1, in defiance of the functionally motivated requirement to the contrary, might be ascribed to this relatively small presence of the hydrosphere in the artist’s mind (for whatever reason). For a more substantive and specific account, see section 4.2 below. Also awaiting explication are the function and alignment of the two inscriptions. Of vital importance about the content of the inscriptions is that neither has a pictorial correspondence at all: the creation of land is not pictured; nor is the sun represented. In this light, one may conjecture, these marginal notes would have been provided to make the existing pictures more intelligible by giving supplementary information needed for a fuller understanding of the whole of the creation pictures.63 Thus, insofar as the creation of land is not illustrated on its own but simply implied by growth of vegetation, the separation of the land and seas is explicitly referred to in the note, well before a switch to P7.2 from P6. Similarly, since the sun is not represented in its transparent form, the creation of the day in distinction from the night (which is depicted in P7.2) needs to be specified on the margin of P7.1, in which the absence of the stars would otherwise lead one to expect the representation of the sun. From the perspective proposed here, then, the particular alignment of the notes to the pictures seems no less appropriate than their function to fulfill.64 63 This conjecture is supported by Lucas’s (1980: 209) credible inference on the collaboration between the scribe and the first artist, as follows: “In Quire 1 [in which P6 and P7 are included: S. S.] what happened was that the scribe wrote the text leaving spaces for the illustrations, the artist drew the illustrations, and the scribe added the captions.” 64 According to Gollancz (1927: xl), the first inscription belongs to P7.1, while the second one should be aligned with P7.2. Gollancz thus had to assume a misplacement of both inscriptions.

4.1 A synchronic–structural perspective 

 101

To elaborate in more specific terms, the two marginal inscriptions would have been intended to make the transition from P6 to P7.2 and from P7.2 to P7.1 conceptually smooth by making explicit some underlying presuppositions not illustrated. Otherwise, one might wonder why plants grow on the surface of waters (since the creation of land is not represented) and what happened to the stars just set in the sky (since the sun is not represented). If the proposed account is acceptable, then, it would be implied that these two notes were provided after P6/P7 were executed: rather than stipulated as directions to the illuminator as assumed by Gollancz and others, they would have postdated the production of the pictures for a better understanding on the part of viewers/readers. A final outstanding feature of the creation pictures is their order in progression: P6 (first and second days) > P7.2 (third and fourth days) > P7.1 (fifth and sixth days). Of focal importance is the bottom–up, rather than top–down, ordering on page 7, namely, P7.2 > P7.1. Therefore, despite the apparent similarity on the surface, the series P6/P7 differs from its predecessor P2/P3 in its narrative development in its latter portion, from top to bottom in P3 on the one hand and from bottom to top in P7 on the other. Of the two alternative sequential orderings, the top-to-bottom seems by far the more natural, as it is in perfect harmony with the textual organization of lines. Yet there are a few instances of the corresponding image arrangement on the vertical axis, as in P41 (Pl. 19; section 10.1), in which the Judgment of the serpent, depicted above, took place subsequent to God calling to Adam and Eve, illustrated below (Genesis A 852a–917b). In any event, the apparently anomalous ordering in P7, especially implemented immediately after the reverse natural progression in P3, would require explanation. Inasmuch as synchronic accounts are unlikely to be forthcoming, one should turn to a diachronic perspective for a proper understanding of the apparently exceptional arrangement of the three pictures. In this connection, it seems revealing to point out some pictorial correlations to the narrative ordering of the three illustrations in question. Of central interest are the gradient ways in which several figures and features are subject to change through the progressing scenes. With respect to the upper arcs, the Creator is shown seated first (P6) and subsequently standing (P7.2 and P7.1); in a correspondingly divergent way, God points downward with his left hand (P6), but he holds a codex in the same hand (P7.2 and P7.1). The Creator is represented beardless earlier (P6 and P7.2) but later bearded (P7.1). In the first two scenes, the light is pouring down over the Creator from the overturned vessel carried by the angel (P6 and P7.2), whereas in the last scene he is alone, encircled by a mandorla (P7.1). In regard to the lower arcs, the first and third panels (P6 an P7.1) are analogously shaped as the semicircles are extended into the rectangular bottom sphere occupied by the water, while the middle register (P7.2) lacks a corresponding water zone below the ground. Thus, the periphery pictures (P6 and P7.1) are shaped in similar fashion in distinction from the central one. In this way, the three pictures are differentiated from each other in multidimensional and intersecting manners by the cluster of pictorial parameters, thereby constituting a gradual transition from one scene to another with some properties unaffected by the advancement.

102 

 4 The Creation: P6/P7

4.2 A diachronic–comparative perspective On the whole, the three creation pictures P6, P7.1, and P7.2 may strike one as unique in their composition and iconography (cf. Raw 1976: 136; Broderick 1978: 112, 137–140). In fact, these images apparently have no directly comparable figures and configurations elsewhere, as substantiated in earlier investigations, and to be confirmed in the following discussion. However, by decomposing them into their constituents—more often than not, elements smaller than are conventionally conceptualized—and reconfiguring them into novel entities, these images exhibit resemblances to the Cotton Genesis tradition at large and its Touronian exemplifications in particular. At stake, in part, are a number of conspicuous features characteristic of P6 and P7 that defy a strictly synchronic/ formal/structural explanation, as the preceding examination has identified. Exploring an explanatory account for them would accordingly have to bring a diachronic/comparative/genealogical perspective to bear on the issues, which in turn would lead to a better understanding of the basically Touronian identities of the Junius creation images at issue. Given that the extant Touronian Bibles do not contain creation pictures of their own, one may well start by examining for comparison other works subsumed under the Cotton Genesis family. In this connection, the Salerno ivories (Bergman 1980: 14–19) and the San Marco mosaics (Weitzmann 1984: 108–111) seem to provide prime comparative materials.65 In overall terms, the San Marco mosaics constitute a more elaborate version of the Cotton Genesis creation cycle than is exemplified at Salerno. At San Marco as many as

65 Needless to say, I would not claim anachronistically that these eleventh- and thirteenth-century artifacts actually provided models for Junius P6/P7. My contention is that Junius would have been indebted to the earlier Cotton sources that are reliably reflected in the Creation cycle at Salerno in the extant corpus. According to recent rethinking, the corresponding cycle in the San Marco mosaics is largely innovative, rather than faithfully testifying to the Cotton Genesis archetype (Kessler 2009: 471; 2009–2010: 21–23; 2014a: 11–14, 16; 2014b: 79, 81; see also section 7.2 below). Moreover, there is credible reason to assume that the San Marco Creation series was inspired by the Salerno plaques (Kessler 2013: 408–410, 414). These considerations may prioritize Salerno over San Marco as a primary comparandum, as will be substantiated independently in the following discussion. Another materialization of the same tradition is an ivory plaque (Berlin, Staatliche Museen zu Berlin, Ident. Nr.: 589; https://id.smb.museum/object/866759/zweiseitig-kreuzigung-christi-und-zehn-szenender-genesis), dated to the late eleventh century, which is carved on both sides, the Crucifixion on one and the Genesis cycle on the other—the series of ten scenes ranging from the first day creation to the expulsion of Adam and Eve (Kessler 1966: figs. 1–2). According to Kessler (1966: 77–86), the creation cycle of the Berlin plaque is closely related to its Salerno counterpart, derivable from the same model affiliated with the Cotton Genesis recension. The Berlin ivory, however, was additionally subject to eastern influence, as witnessed notably by the head of Abyssus and the Creator in the medallion, both occurring in the initial panel for the first day creation (Kessler 1966: 79), but lacking at Salerno and San Marco. Moreover, the Berlin plaque would have been damaged in the course of production, which may be held responsible for the incomplete set of scenes in the extant work. In view of these interferences, one may legitimately treat the Berlin plaque as of secondary importance for present concerns.

4.2 A diachronic–comparative perspective 

 103

nine panels are devoted to the creation scenes extending to the creation of animals, referred to as SM 1 through SM 9 in Table 9 below (Demus 1984: pls. 108–115; Büchsel/ Kessler/Müller 2014: pls. 5–11), whereas no more than six panels are allotted to the same cycle at Salerno (Pls. 52–55). Thus, half the six panels at Salerno—assigned to the first, third, and fifth days—have two corresponding zones at San Marco, and the other half at Salerno—corresponding to the remaining three even-numbered days of creation—are each matched with a single zone at San Marco. Bringing in Junius 11 with still fewer creation pictures as examined in the preceding section, one comes up with the following correspondence table of the Creation cycle in Junius, Salerno, and San Marco (Table 9). Table 9. Correspondence table of the Creation cycle in Junius, Salerno, and San Marco Day

Junius

Salerno

San Marco

1st 2nd 3rd 4th 5th 6th

P6

Sal 1.1 Sal 1.2 Sal 2.1 Sal 2.2 Sal 3.1 Sal 3.2

SM 1+2 SM 3 SM 4+5 SM 6 SM 7+8 SM 9

P7.2 P7.1

Key: Sal 1.1: Creation of Light and Darkness; Pl. 52 (left); Dell’Acqua et al. 2016: pl. 1a. Sal 1.2: Creation of the Firmament; Pl. 52 (right); Dell’Acqua et al. 2016: pl. 1a. Sal 2.1: Creation of the Plants and Trees; Pl. 53 (left); Dell’Acqua et al. 2016: pl. 2a. Sal 2.2: Creation of the Sun, the Moon, and the Stars; Pl. 53 (right); Dell’Acqua et al. 2016: pl. 2a. Sal 3.1: Creation of Birds and Fish; Pl. 54; Dell’Acqua et al. 2016: pl. 3a. Sal 3.2: Creation of Animals; Pl. 55; Dell’Acqua et al. 2016: pl. 3c. SM 1: The Spirit above the Waters; Dumbarton Oaks Research Library and Collection SM0748; Demus 1984: pl. 108; Büchsel/Kessler/Müller 2014: pl. 5. SM 2: God Separating the Light from the Darkness; Dumbarton Oaks Research Library and Collection SM0749; Demus 1984: pl. 109; Büchsel/Kessler/Müller 2014: pl. 6 (left). SM 3: God Creating the Firmament; Dumbarton Oaks Research Library and Collection SM0752; Demus 1984: pl. 110; Büchsel/Kessler/Müller 2014: pl. 6 (right). SM 4: God Separating the Seas from the Dry Land; Dumbarton Oaks Research Library and Collection SM0754; Demus 1984: pl. 111; Büchsel/Kessler/Müller 2014: pl. 7 (left). SM 5: God Creating the Plants; Dumbarton Oaks Research Library and Collection SM0756; Demus 1984: pl. 112; Büchsel/Kessler/Müller 2014: pl. 7 (right). SM 6: God Creating the Heavenly Bodies; Dumbarton Oaks Research Library and Collection SM0758; Demus 1984: pl. 113; Büchsel/Kessler/Müller 2014: pl. 8. SM 7: Creation of the Birds and Marine Creatures; Dumbarton Oaks Research Library and Collection SM0763; unillustrated in Demus 1984; Büchsel/Kessler/Müller 2014: pl. 9. SM 8: God Blessing the Birds and Marine Creatures; Dumbarton Oaks Research Library and Collection SM0763.1; Demus 1984: pl. 114; Büchsel/Kessler/Müller 2014: pl. 10. SM 9: God Creating the Terrestrial Animals; Dumbarton Oaks Research Library and Collection SM0765; Demus 1984: pl. 115; Büchsel/Kessler/Müller 2014: pl. 11.

104 

 4 The Creation: P6/P7

For the obvious reason of limited space available, the Salerno plaques appear closer to Junius in regard to the scope and manner of visualization of the creation story. For instance, the primeval waters are represented in conjunction with the creation (or separation) of light and darkness in the same panel, commonly in Junius (P6) and Salerno (Sal 1.1, Pl.  52), whereas these two scenes are represented separately at San Marco (SM 1/2; Demus 1984: pls. 108, 109). Unlike San Marco, the water is not divided into heaven and earth in Junius and Salerno. Moreover, the separation of water and land is not depicted in Junius and Salerno, in distinction from San Marco (SM 4; Demus 1984: pl. 111). By far more interesting are the parallel ways in which the corresponding textual materials are organized and arranged for pictorial representation at Salerno and Junius. That is, in essential respects, the Salerno ivories exhibit greater similarities in overall organization to Junius 11 than do the San Marco mosaics in representing God’s creation. God’s works of creation up to but not including the creation of Eve are represented in three two-paneled plaques in the Salerno ivories: the first plaque (Sal 1, Pl. 52) depicts the creation of light and darkness on the left side and the creation of the firmament on the right;66 the second plaque (Sal 2, Pl. 53) is devoted to the creation of plants on the left and that of the sun, the moon, and the stars on the right; and the third plaque, originally composed of two panels but now existing in two distinct pieces, is concerned with the creation of birds and fishes on the one hand (Sal 3.1, Pl. 54) and of animals on the other (Sal 3.2, Pl. 55). There are a number of notable similarities, and no less noteworthy dissimilarities, between the two works. First, the tripartite structure of the Salerno plaques, with each plaque allocated to two days of work as mentioned above, is similar to Junius’s threepart organization. This structural parallelism ends here, however. Conspicuously different is that, while the two–day works are each highly integrated into a single representation in Junius, at Salerno, one day is represented in its own panel independent of and sharply delineated from the other paired with it. In short, a conflation on the one hand and a juxtaposition on the other are in force in Junius and Salerno, respectively, of two days of works. Second, internal to each plaque, it is bipartite, much as each picture of Junius (P6, P7.1, P7.2). However, while in Junius only one of the two components in each of the three pictures is allocated to earth and hence to an integrated scene of creation for two successive days, at Salerno the two constituent panels in each plaque equally represent separate scenes of the Creator’s work on two consecutive days. The independence of each day’s work is indicated by the presence of God to the immediate left of the event of creation being depicted, as further noted below. Third, one of the two components is stacked vertically on the other in Junius, whereas the two constituent panels are placed in a horizontal sequence in the Salerno plaques. Fourth, derived from the second point 66 The omission of God originally at the left in the first scene (Sal 1.1) and the absence of the disc originally to be found between the Creator and the angels (Sal 1.2), in contrast to their presence at San Marco (SM 2/3), may have been related. Both would have been found redundant given their comparable figures in the other panel.

4.2 A diachronic–comparative perspective 

 105

raised above, the two concatenated panels progress in temporal succession from left to right in the Salerno creation cycle, whereas the two vertical tiers are presented as an integrated synchronic state without an internal division in time. For example, the separation of light and darkness on the one hand and the raising of the firmament on the other are illustrated in separate panels at Salerno. In contrast, the light and the darkness, personified in the form of angels, belong to the two different discs in Junius, while the circumference of the firmament constitutes the border between the two zones. Fifth, while the Creator is situated above the created in Junius (with the exception of light in P6), he stands to the left of his creations at Salerno (with the exception of the first day of creation where God is not depicted). In other words, the Creator and the created stand face to face on the same ground in the Salerno plaques. By contrast, in the Junius cycle, God is detached from the scene of creation, beholding from on high. Despite the conspicuous differences in the pictorial organizations between Junius 11 and the Salerno plaques, as pointed out above, the Junius system is derivable from the Salerno counterpart on the basis of a chain of well-designed and motivated transformations. Of the first three two–paneled plaques of the Salerno ivories representing the six days of creation by coupling two–day scenes side by side, the initial two pieces happen to be arranged in such a way that the temporal order corresponds to the spatial distinction in the vertical dimension. Specifically, while the first (left) zone of the first two plaques (Sal 1.1 and Sal 2.1; Table 9) illustrates primarily an event down on earth, the second (right; Sal 1.2 and Sal 2.2; Table 9) one is concerned with a scene up above the earth. In order to coordinate such temporal and spatial correspondences optimally and most naturally in cognitive terms, the original left-to-right orientation (linear order) as at Salerno would have been reorganized along the vertical orientation, resulting in the bottom-to-top orientation as in Junius. Of relevance in this connection is the vertical orientation of the two initial images of God enthroned, namely, Pii and P2 (section 2.1). Preceding the Creation, both scenes leave the lower half void, as described in Genesis A 103a–110a: Ne wæs hēr þā gīet / nymþe heolstersceado // wiht geworden / ac þes wīda grund // stōd dēop and dim, / drihtne fremde, // īdel and unnȳt, / on þone ēagum wlāt // stīðfrihþ cining / and þā stōwe behēold, // drēama lēase, / geseah deorc gesweorc // sēmian sinnihte, / sweart under roderum, // wonn and weste (Then was nothing yet here except darkness, but this vast abyss stood deep and dark, alien to the Lord, idle and useless. The resolute king looked upon it with his eyes, and beheld the place, without joys, saw the dark mist hanging in perpetual night, black under the skies, gloomy and void.) (Anlezark’s [2011: 9, 11] translation)

In this light, the reconstitution of the Creation cycle along the vertical axis in P6/P7 may be viewed as a materialization of the empty space provided in Pii/P2. Thus, the composition of P6/P7 would have been motivated in no small measure by the logic of pictorial representation inherent in Junius as inscribed in the first two pictures. The trio of P6/ P7.1/P7.2 fill in the empty space—the lower half domain—with concrete creations, while the upper half—the celestial sphere—is depicted in simplified form in comparison with

106 

 4 The Creation: P6/P7

the original exemplars, the pair of Pii/P2. Also responsible would have been the originally vertical orientation of other plaques that were included in the Salerno ivories, such as the plaques given in Bergman (1980: fig. 20) and many others that follow. Concomitant with the verticalization of the two constituent scenes, however, they become susceptible to integration with each other in a single canvas, whereby the sense of their synchronicity comes to override that of their sequentiality. As a consequence, the categorical demarcation between the two panels vertically placed loses its compositional significance, thereby formally accomplishing a full integration of the originally distinct picture zones. The above reconceptualization in the domain of an individual plaque would have brought in its train no small consequences for the relation beyond the plaque boundary. More specifically, since the right side corresponds to the upper domain (heaven) at Salerno and the like, and since the right side follows the left in temporal sequencing, it would be feasible to extend this relation beyond a given single plaque to a larger domain of adjacent plaques. As a consequence, in the transformation from horizontal to vertical placement, the first (left) panel of a following plaque is placed above the second (right) panel of the preceding one, so that the consistent order from bottom to top comes into effect both within and without an individual plaque, that is, exactly the arrangement that was implemented on page 7 of Junius in its entirety. From a purely synchronic perspective, there seem to be no intrinsic reasons for preferring the current order (from bottom up) to the otherwise natural alternative (from top down, as in P3). In fact, the pattern of textual organization would obviously favor the top-to-bottom progression, as observed at the end of the preceding section. Therefore, one may have no choice but to explore a diachronic account of this synchronically puzzling situation, which has now been resolved at least in part, as addressed in the preceding paragraph: the bottom–up order is determined as a structural consequence of the reorientation of constituent panels from the original horizontal to the vertical axis. Yet there is a further dimension to the issue of bottom–up order for exploration. At this juncture, Raw’s (1976: 136) postulation of a triptych as the origin of P6/P7 (mentioned in section 4.1 above) proves to be illuminating when seen in a different light than she originally conceived. Unless special requirements had been in force to obviate optimal conditions for pictorial representation, the tripartite pictures resulting from the reorganization along the vertical axis, as proposed above, would most naturally have been realized as full-page illustrations each. Moreover, if canon tables had served as part of the analogical bases for the reorganization of the model as fully argued below, there would have been a further motivation for such a three-page sequencing of double– arc constructs. Accordingly, P6, P7.1, and P7.2 would each have occurred on separate pages. Such optimal full-page illustrations, however, would have been compromised at some point in the history of manuscript transmission. The question of when the change from the three-page to the current two-page scheme took place is not of primary importance here: of overriding interest rather is the presumable way in which the two-page arrangement P6/P7 came to be implemented by reorganization, whether directly from

4.2 A diachronic–comparative perspective 

 107

the alleged model of an ivory triptych (cf. Raw 1976) or derivatively from the three-page allotment in an antecedent manuscript. In any event, regardless of the actual materialization, the three full-page pictures faithful to their original execution would have been conceptualized as an inseparable whole that was organized in such a way that the three component pictures are strictly ordered, A–B–C (A = P6, B = P7.2, C = P7.1), whereby the linear–temporal adjacency matches the spatial proximity. This optimal sequence would have been subject to restructuring at some point in transmission when the three-page distribution had to be replaced by the two-page one, probably dictated by allocation of space for the running text. Confronted with the new requirement, one would then have had the following eight possible arrangements at his disposal, under the general condition that one follows for text (as opposed to pictorial) comprehension a top-to-bottom processing across lines within a single page first and then a left-to-right processing across pages (Key: A, B, C = pictures, as specified above; X = text; 6, 7 = pages):67 a

 

 

b 

 

 

c 

 

 

d 

 

 

e 

 

 

f 

 

 

g 

h 

A 

C   

A 

X   

B 

C   

B 

X   

A 

B   

A 

C   

X 

B 

X 

C 

B 

X   

B 

C   

A 

X   

A 

C   

X 

C   

X 

B   

A 

C 

A 

B 

6 

7   

6 

7   

6 

7   

6 

7   

6 

7   

6 

7   

6 

7 

6 

7 

Among the eight options listed above, the first two, (a) and (b), are favored over the second two, (c) and (d), on account of the order involved that adheres to the natural top-to-bottom arrangement. With respect to these preferable two, (b) may count as less appropriate than (a), because pictures B and C are intervened by an extraneous element (i.e., the text) and thus have to be treated as discontinuous relative to the textual flow, although their intrinsic sequentiality remains unaltered. Similar considerations may apply to the last four options: the disruption by text will render arrangements (e) and (f) less suitable for an integral representation. Thus, options (g) and (h) remain to be evaluated, in addition to (a). Choosing between the former two, it is (h) that corresponds more closely to the immediate proximity in time and space distinctively characteristic of the original, whereby A is directly followed by B in spatial arrangement as well as narrative unfolding. Specifically, the distance between A and B is larger than that between A and C in (g) contrary to the narrative proximity involved, whereas the spatial and narrative distances are analogously valued in (h). Accordingly, the maximal fidelity to and hence the minimal distortion of

67 Needless to say, the numbering of 6 and 7 is biased to Junius 11 in its extant form. In more general terms, any pair of consecutive numbers would do, insofar as the first member is an even number to insure that the whole cycle may be presented at a glance in order to give full justice to the pictorial integrity involved.

108 

 4 The Creation: P6/P7

the original representation—the pictorial integrity and the internal sequential consistency—select (a) and (h) as finalists to be chosen from. These two in turn have their own advantages and disadvantages. While (a) suffers from the spatial incongruence between B and C that fails to match the temporal and the narrative adjacency, (h) contravenes the otherwise prevailing top-to-bottom directionality in textual and pictorial processing. Making a final decision on arrangement must accordingly rest on external weighing beyond strictly pictorial issues. At stake is the allocation of textual space: should the half-page textual material be placed before or after the pictures, namely in the first half of page 6 (h) or the second half of page 7 (a)? It deserves particular attention that the reduction by half of the presumably original three-page illustrations may ultimately be attributed to the apportioning of a half page for textual use, which would hardly make the optimal full-page representation viable. In view of such a text–based motivation, letting textual considerations dictate the final decision on the choice of pictorial arrangement should be appreciated as reasonable thinking. The relevant portion of the text occurring on page 6 in Junius 11 ranges from Genesis A 114b through 135b. In exploring the ways in which the text interfaces with the pictures, there are two significant dimensions—internal and external—to be addressed in conjunction. Internally, how does the text connect with its visualizations in terms of subject matter? Externally, how does the text articulate with the pictures in its outer form? From the perspective of narrative development, the two competing placements— at the beginning of page 6 or in the middle of page 7—differ in their references to the pictures. Continuing from the preceding page, the passage in question narrates the eternal darkness existing before creation and subsequently God’s work of creation on the first day, that is, the separation of light from darkness and their naming Day and Night. Placed on page 6 before the initial panel as in the extant manuscript, the text and picture, both initiating their own series of representation—verbal and pictorial—fit each other optimally, the two separate vehicles of information—text and picture—situated analogously and iconically on the same page. By contrast, if ordered after the third picture (P7.1) on page 7, the text must refer far back to the initial illustration, as a consequence of the corresponding text and picture located at the polar opposites of the two-page space. As far as the external articulation between text and picture is concerned, there are the two stochastic generalizations in effect, as identified in Chapter 1. First, a picture may be preceded but not followed by the text on the same page. To reformulate, a picture is optimally aligned with the bottom of a page. This requirement for bottom alignment chooses (h) over (a), as the latter allows the text to follow the picture C. Second, and more specifically, a picture is more likely to be inserted after the b-verse and before the a-verse. In other words, the line break, a higher metrical unit, is more tolerant of disruption by pictorial material than the verse break, a lower unit. In this light, the extant location of the text in the first half of page 6 before the pictorial cycle begins, that is (h), obeys the rule: the initial picture A is placed after verse 135b. By contrast, the alternative

4.2 A diachronic–comparative perspective 

 109

arrangement after the P7.1 picture in the middle of page 7, that is (a), would result in the picture cycle being inserted after the a-verse (114a) and before the b-verse (114b), thereby breaching the generalization. Therefore, on both grounds—picture alignment with page bottom and picture placement between the b-verse and the a-verse—the allocation of the first half of page 6 to the textual body counts as more favorable than the alternative solution, as it creates an optimal articulation between text and picture. The foregoing discussion provides independent motivations for allocating the text in the first half of page 6. This well-motivated decision, then, singles out (h) as the optimal arrangement to adopt. Therefore, the second picture of the creation cycle in Junius occupies the second half of page 7, below the final one of the series, as one actually sees in the manuscript. Meanwhile, the original square framing of each panel undergoes reshaping to be reorganized on the basis of an arc-shaped framework. The transformation from rectangularity to circularity in overall shape would have been motivated on several converging grounds. First, the prior existence of circular objects represented in the models would have naturally served as sources of inspiration: the light and the darkness (the first day, both Salerno [Pl. 52] and San Marco [Büchsel/Kessler/Müller 2014: pl. 6 left]), the sun and the moon, and, most importantly, the firmament on which the celestial bodies are set (the second day, San Marco [Büchsel/Kessler/Müller 2014: pl. 6 right]; the fourth day, both Salerno [Pl. 53] and San Marco [Büchsel/Kessler/Müller 2014: pl. 8]). Second, the relocation of the Creator high in the upper zone above the created in the initial plaque would have invoked by association the familiar image of God in Majesty, particularly God seated on a globe (conceptualized as the firmament, corresponding to the lower dome) and enclosed in a mandorla (corresponding to the outer dome), that is, the figure of God represented within a frame of double or multiple nested circles, as exemplified in Prüm Gospels (Berlin, Staatsbibliothek zu Berlin, Ms. theol. lat. fol. 733), fol. 17v (Köhler 1930: pl. I.93c; Kessler 1977: fig. 52); Grandval 352v (Pl. 25); Vivian 329v (Pl. 30); San Paolo 259v (Pl. 36); San Paolo 260v, 270v, 277v, 287v; Lothair Gospels (Paris, Bibliothèque Nationale, Ms. lat. 266), fol. 2v (Köhler 1930: pl. I.98b; Kessler 1977: fig. 61); Dufay Gospels (Paris, Bibliothèque Nationale, Ms. lat. 9385), fol. 179v (Köhler 1930: pl. I.109a; Kessler 1977: fig. 62); Le Mans Gospels (Paris, Bibliothèque Nationale, Ms. lat. 261), fol. 18r (Köhler 1930: pl. I.118b; Kessler 1977: fig. 63); and many others. The image of God in Majesty expands by further metaphorical association to an invocation of the King on the throne under an arch, on the one hand, and the Evangelist at work under an arch, on the other, with all their attendant architectural and other figures (for details, see below), as exemplified in Vivian 423r and Dufay Gospels 18v (Köhler 1930: pl. I.110a)/Prüm Gospels 22v (Köhler 1930: pl. I.94), respectively.68 Another line of analogical imaging, induced by the presence of elaborated pillars in the models, as well as by the representation of the

68 Outside the School of Tours, see, for example, the Gospels of St. Médard de Soissons (Paris, Bibliothèque Nationale, Ms. lat. 8850), fol. 81v (Mütherich/Gaehde 1976: pl. 6).

110 

 4 The Creation: P6/P7

king or the evangelist under the arch, would have conjured up the canon tables topped with arches and embellished with similar architectural details. The resultant vertical reorientation and the consequential double-arch framing in their turn would have led to radical reconfigurations and recompositions of constituent figures contained therein, including their subcomponents. All the major motivations and materials for adaptation and reworking would have been available internally within the models and their immediate contexts by virtue of the network of associations, conceptual, iconographical, and compositional, as substantiated in the following examinations of each of the three creation drawings in Junius 11 (P6, P7.1, and P7.2). Starting with P6, upon verticalization of the two panels, the light and the darkness— imaged in circular form (Salerno 1.1, Pl. 52, left)—would likely have been polarized in spatial distribution in harmony with the innovated orientation, whereby the light is relocated to the upper panel. One of the four angels depicted in the right zone (Salerno 1.2, Pl. 52)—most accessible would have been the one standing in front on the viewer’s side—is then chosen and placed directly above the disc of light, now relocated in the same upper zone of P6. The resulting configuration of the angel stretching both arms downward to the disc turns out to be strikingly similar to the figure of the angel holding a bowl of light in P6. By further refinement and reconfiguration (with necessary readjustments in size and pose, notably from standing to a hovering posture; see further below), the two coordinated figures are transformed into the angel holding the vessel of light in Junius P6. This implementation of recomposition and reconfiguration constitutes the essential component of the personification of light distinctively characteristic of Junius 11.69 Thus, one of the four angels lifting the firmament at Salerno is reinterpreted as the one holding the source of light, which is now visualized most concretely as a bowl of light in Junius, an item that would have been found more appropriate for holding in cognitive terms. Apart from a general cognitive basis, the co-representation of the source of light as a vessel specifically would have been motivated by metaphor and metonymy. As argued above, the arch–based architectural structure would have been owing in part to the representation of the King on the throne on the one hand, and the canon tables on the other. More often than not, vessels/lamps/chalices are shown hanging from the top 69 As is usual with the Utrecht Psalter, the image of an angel taking away the moon (Utrecht 40v, Pl. 77) might have played only an anecdotal role in motivating the composition of the angel holding a vessel of light (cf. Broderick 1978: 115, 139). Given the conceptual remoteness and the compositional/situational dissimilarity involved, the Utrecht image would unlikely have given rise to the novel configuration at issue in its own right, despite Broderick’s (1978: 115) view to the contrary. Specifically, the image in question is conceptually far removed from lightening coming down from above to the world below: the angel is moving away, rather than staying still above earth; contrary to the conceptualization in force in Utrecht 40v, the moon is closely associated with darkness, the night, in P7.2, as discussed below; and the angel hardly can be conceptualized as a representative of light in the Utrecht image. What would have been involved rather is simply the superficial combination of an angel and a circular form of light in a single image.

4.2 A diachronic–comparative perspective 

 111

arch in such images (e.g., Vivian 423r; cf. Diebold 1990: 104–106; Garipzanov 2008: 245). Furthermore, located in the analogous position—with or without such dishes—God’s hand is often represented emitting downward rays of light (e.g., Vivian 423r; cf. Diebold 1990: 107–114; Garipzanov 2008: 244–245). The clustering of these images under the arch most likely would have given inspiration to the illustration of the largely comparable figures in the corresponding place, namely, the angel holding a vessel of light from the top of the arch. The choice of just a single angel at the sacrifice of the other three would have been facilitated by a similar composition at San Marco,70 in which the number of angels depicted stands for the days of creation, a convention not followed at Salerno and Junius (Broderick 1978: 138; Bergman 1980: 15). The single angel symbolizing the first day standing just behind the source of light—the disc—at San Marco would have given further impetus to the reanalysis whereby the angel in question was inseparably associated with the light, thereby leading the Junius artist (or his predecessor) to reconceptualize the angel as integral to the creation and separation of light (rather than a bystander of the creation scene), namely, as the figure carrying the vessel of light. This chain of reconfiguration and recomposition may have been further inspired at the outset by the personified representations of the sun and the moon, at both Salerno and San Marco. Thus ultimately responsible for the personification of light would have been the identification of light and the sun as covariants of each other. The postulation of such an identity relation may be ascertained in the structural parallelism between light/darkness and the sun/the moon at San Marco and Salerno. The first member of each pair is placed closer to God than the second one, self-evidently at San Marco. Given that God stands invariably at the left side of the panel in all of those in which he is depicted at Salerno, the appearance of light on the left in the first panel (Salerno 1.1, Pl. 52) indirectly confirms the generalization. Accordingly, such a formal correspondence may be adduced as a reflection of the two pairs’ intimate relation, a cognitive basis for going further to subsume the two sets of opposites under the same superordinate category. Subsequently, through opposition to the light, the darkness would have been subject to a comparable personification, a transfiguration that eventually replaced the original spirit in the form of a dove (as at Salerno and San Marco) with an angelic being. It is worth emphasizing that, despite the substitution, the overall shape of the figure is largely retained. Of conspicuous similarity are the spread wings and the drapery that flows to the right downward in P6, reminiscent of the corresponding dove’s body and tail as represented at Salerno. Concomitant with this process of personification, the initial embodiment of darkness, that is, the right circle in the Salerno plaque, came to be deprived of its original signifying power, losing its autonomous significance as a

70 For a less articulated and elaborated suggestion of a pictorial transformation from the San Marco image to the Junius one of the angel carrying a bowl of light, see Blum (1976: 217), which duly prompted Broderick’s (1978: 140n71) objection.

112 

 4 The Creation: P6/P7

representable entity, and thereupon dissolving into the primeval waters underneath. As a consequence, the darkness is now embodied entirely by the configuring of the angelic symbol above and the water below. As with P6, P7.2 underwent a reorientation along the vertical axis, thereby having the right half of Salerno’s second plaque stacked on top of the left counterpart. In the course of verticalization, the two originally separate panels are integrated into a single whole space. This spatial conflation in turn obliterates the temporal distinction between the two scenes, and merges the two days of creation into a common synchronic state. Although the two panels involved each illustrate God, the resultant P7.2 contains a figure of the Creator only in the upper zone, following the pattern set by the initial scene in P6. In addition to the major difference in orientation, P7.2 exhibits remarkable modifications of the model, partly induced by the prior composition in P6 and partly introduced anew in their own right. Of particular interest are, first, the arrangement of the heavenly bodies, and second, the representation of the plants. While the sun, the moon, and the stars are all depicted and distributed evenly in the circle at Salerno and San Marco, only the stars are apparently illustrated in Junius, and are limited to the upper periphery of the lower disc. At first glance, then, Junius seems to exhibit no resemblance at all to the Cotton Genesis tradition in this respect. At a deeper level of organization, however, P7.2 proves to be analogous to Salerno and San Marco with varying degrees of verisimilitude. The placement of the sun above the moon and the stars at San Marco is partly retained in Junius, in which what may be conceived as the sun, through identification with the source of light depicted in P6 (for details, see below), is positioned in the upper disc, while the stars are set on the lower dome, or restated in alternate terms, they occupy the lower periphery of the upper disc overlapping the upper zone of the lower disc. This arrangement of the sun at the top at San Marco contrasts with Salerno, in which the sun and the moon are set side by side on the same level. This would by no means imply that the privileged placement of the sun on top is indebted solely to San Marco’s composition. At Salerno, while on the same vertical level, the sun is placed horizontally closer to the Creator than the moon. This immediate proximity of the sun to God, the configuration that commonly obtains at San Marco and Salerno in an abstract relative sense, would then have been radically reorganized in Junius in absolutely vertical terms, in harmony with the overall reorientation along that dimension characteristic of P6 and P7 in this manuscript. Accordingly, the sun would have been placed at the top and the moon at the bottom of the upper dome, had it not been for other factors intervening in their full materialization. Because of the partial overlap of the two arcs, the moon would likely have been positioned at the top of the lower dome, as the stars are in actuality.71 The sun’s vertical relocation would have coincided with the corresponding space for the angel holding the source of light in P6. This spatial coincidence, in conjunction with the common essence of light involved, would

71 The actual absence of the moon in the resultant picture will be explained in due course.

4.2 A diachronic–comparative perspective 

 113

have motivated the identification of the two entities—the source of universal light and the sun—sharing the analogous place, thereby resulting in the replication in P7.2 of the same symbol—the light-bearing angel—in agreement with P6. Thus, the sun, now symbolized by the bowl of light held by the angel upon reidentification, is shown hanging over God’s head, in polar opposition to the remaining heavenly bodies (except the moon) found under God’s feet. Of related interest in this connection is that the star positioned at the top of the constellation is distinguished in shape and size from the others both in Junius and at Salerno, whereas no comparable distinction is made at San Marco. Drawing on the assumption that the angel holding the bowl of light stands for the sun in P7.2, as suggested by Broderick (1978: 124), the foregoing argument has shown how, in the course of derivation from the model like the Salerno plaque, the sun, created on the fourth day, came to be associated with the same symbol originally in use for the light brought into being on the first day of creation. One may legitimately pose a fundamental question at this point: Reservations expressed in the preceding section aside, regarding the identification of the light-bearing angel with the sun on synchronic grounds, is there evidence from a complementary perspective for identifying the second occurrence of the angel (P7.2) specifically as the sun, rather than a simple carry-over effect from P6, as with the firmament created on the second day? Actually, there is reason to interpret the representation as the sun, at the expense of its negation, “the angel does not stand for the sun.” At issue is the posture of the personified sun at Salerno—especially the gesture of its hands and arms—which is strikingly similar to the angel holding the vessel in Junius, as a rotation of the sun at Salerno by 90 degrees will readily show. This may suggest that the construction of the symbol for light itself in Junius is owing to the symbolization of the sun as at Salerno. Thus, the two kinds of light, light as such—the first day creation—and the sun—its fourth day counterpart—would have been inseparably conceived from the start in the mind of the Junius artist. Interpreted this way, the angel holding a vessel of light may well be identified as a sign vehicle for light in general, defying a categorical division between universal source of light and its celestial manifestation, the sun. Accordingly, rectifying the earlier flat rejection of such identification requires modifying the assessment as follows: the angel holding a vessel of light may as likely signify the sun as not, given the inherent ambivalence of the symbol concerned. One may suspect that the artist seems to have been content with leaving the bowl deliberately ambiguous in its denotation. Under the circumstances, then, to assert the reverse as assumed in the previous section, “the sun is not represented at all in P7.2,” is too strong an assertion to be accepted. As highlighted in the above argument, the Salerno ivory plaque (or its close cognate) would have been no less instrumental in shaping the P7.2 image than the San Marco mosaics, despite an apparently more outstanding difference between Junius and Salerno in the way of arranging and representing the celestial bodies. In this connection, the plausibility now revealed—that the personification of the sun at Salerno motivated, in part, a unique representation of the light-carrying angel in Junius—is of paramount interest.

114 

 4 The Creation: P6/P7

While the recomposing and repositioning of the sun and the stars receive credible accounts, the apparent absence of the moon in P7.2 still requires explanation. As argued above, light and darkness were personified in P6, presumably induced by their identification, and hence their conflation, with the sun and the moon, respectively. This personification of the light and darkness would in turn have been motivated by the personified representation of the sun and the moon at Salerno and San Marco. As reasoned in the foregoing discussion, the lower disc of P6 symbolizes the darkness in the form of the angel covering his face and the waters underneath, while the bowl of light held by the angel is found in the upper disc, open, overturned, and radiating. To generalize, while the light is explicitly represented in terms of the presence of positive features, pouring rays of light, the darkness is expressed negatively by the absence of light: compare Genesis A 124b–125a “him was hālig lēoht // ofer wēstenne” (holy light was made by God over the void). Thus, the darkness is conceptualized as primeval background over which the light is made and from which it is separated. Taking this perspective to its logical conclusion, the night, embodied in the lower disc as witnessed by the stars in the sky, may well be represented by the absence of its primary symbol, the moon, the covered moon corresponding to the angel with the veiled face in P6. Failing to be manifested in P7.2, the moon makes its hidden existence recognizable by implication (cf. Broderick 1978: 124): driven out to the lower periphery of the upper disc through polarization to the sun (as discussed above), the moon is covered by the overlapping portion of the lower disc at the intersection between the two stacked arches. Interpreted this way, the opposition between the upper and lower discs is analogous in P6 and P7.2, namely, light versus darkness, as Raw (1955: 318; 1976: 136–137) pointed out from a different perspective. Regarding the plants as depicted in P7.2, they conspicuously differ from those illustrated at San Marco and Salerno. While in the latter two works the plants are shown mature and prospering with flowers and fruits, their Junius counterparts appear bare, feeble, and immature in stature. Rather, they look highly similar to those represented in the Ebbo Gospels (Épernay, Bibliothèque Municipale, Ms. 1), fol. 18v (Mütherich/Gaehde 1976: pl. 14), and, more remarkably, the Codex Aureus of St. Emmeram (Munich, Staatsbibliothek, Clm 14000), fol. 16r (Mütherich/Gaehde 1976: pl. 35). Originated from Reims and appropriated by the Court School of Charles the Bald, these plants, growing thinly on the wavy, cloud-like ground under the arch, eloquently testify to Reimsian influences (cf. Mütherich/Gaehde 1976: 102). Of no less importance is the setting in which these Reimsian plants are represented: as part of an Evangelist portrait, on the cloud/ ground and under the arch, with the intervening space finished like the sky in shape and color but without celestial bodies. That is, the whole configuration involved seems to prefigure the landscape coming into being in the lower disc of P7.2. This outstanding agreement in composition and configuration may provide cogent support for the thesis proposed above that the creation pictures in Junius 11 were inspired, in no small measure, by the images of the King on the throne and of the Evangelist at work in an architectural setting.

4.2 A diachronic–comparative perspective 

 115

P7.1 exhibits its own distinct features awaiting diachronic–genealogical explanations. First, the figure of God differs from the two earlier representations, P6 and P7.2. He is shown standing alone, unaccompanied overhead by the angel holding a bowl of light. Since the third pair, consisting of the fifth and sixth days, is exclusively concerned with earth, it is most unlikely that the verticalization, as implemented on the two preceding sets of illustrations, would transform the right-side panel in the model into an upper sphere in a way corresponding to the two earlier ones. Thus, while conforming on the whole to the overall vertical reorientation, the new upper panel cannot inherit any features from the model (such as animals; cf. Pl. 55), except the standing figure of the Creator, who would naturally have been relocated from earth to heaven. Accordingly, the new upper zone—including its lower periphery overlapping the lower zone—comes to be represented in its neutral blank form without any celestial bodies, particularly the angel holding the bowl of light, that is, the personification of light, and by extension, of the sun. In the absence of other images to model after, then, there is no viable alternative to showing God standing alone and enclosed in a mandorla, a neutral, unmarked representation of God widely practiced. It should be noted in this connection that God is not flanked by any angelic beings, despite their usual attendance. This absence may be attributed to the analogous situation at Salerno, the closest approximation in existence to the model of Junius, as far as the creation cycle is concerned. In this light, standing far away from the scene of creation on earth, God may be seen as not so much engaged in creation as beholding and blessing the created. This reinterpretation of God’s engagement seems to be in full harmony with the appearance of plants as well as the animals in P7.1, as will be shown below. As may be recalled in this connection, since the illustration in the lower arch is exclusively concerned with the accomplishments of creation on earth, whatever was created in heaven is omitted from representation, precisely as at Salerno and San Marco. Like P7.2, P7.1 is distinguished by a number of peculiarities in the ways of representing the creatures brought into being on the fifth and sixth days of creation. Although God created birds and fish on the fifth day, the fish fail to be depicted in the Junius picture, in contrast to Salerno and San Marco, in both of which the two species are represented separately in the upper and lower halves of the same panel. Moreover, only one bird and one beast are illustrated in Junius, whereas a large group of each are found at San Marco and Salerno. In this regard, it may be interesting to compare with P11 and P13, in which groups of animals are illustrated, even a shoal of fish in the case of P13, in ways reminiscent of the San Marco and Salerno pictures. One may be led to suspect accordingly that the artist had at his disposal a corresponding composition to execute for P7.1 as well, had he chosen to do so. There would have to be a good reason to abstain from that option. Finally, unlike the bird and the animal, plants are depicted in multiples as part of the fifth- and sixth-day creation scene in P7.1. By contrast, plants do not appear at Salerno at all, a total omission that seems reasonable, given that they are not created on the fifth and sixth days but well before on the third. The situation at San Marco looks somewhat ambiguous: indeed the ground is painted green in the scene

116 

 4 The Creation: P6/P7

of the creation of animals, suggesting that the land is covered with grass, but there are no trees specifically illustrated in contrast to the third–day scene. As argued above in regard to P6 and P7.2, the double–dome framing executed pervasively in the creation pictures in Junius 11 is indebted to the architectural frame that is extensively employed in the canon tables, the representations of the King on the throne, and the Evangelist portraits. More often than not, such wide-ranging representations exhibit on or under the top arch a frieze of figures comprising plants, birds, animals, and/or humans, as exemplified in the Codex Aureus of St. Emmeram 16r referred to above. This inspiration originating from the arch–based representation would by entailment have imposed an adherence to a frieze or one–dimensional format (cf. Suzuki 2000: 15–17), on the part of the Junius artist (or his predecessors) in adapting and reorganizing the dome structure for the creation cycle. This would have constituted a basis for many, if not all, of the major distinguishing properties of P7.1 mentioned above. Had fish been depicted in the lower disc of P7.1 in full conformity with the text, they would have appeared in the water beneath the frieze of plants and animals, but such a configuration would have resulted in a two–dimensional representation in which figures are illustrated extending along two axes, not only horizontal but also vertical. Given that the models allow for only a horizontally expanding, one–dimensional space for visualization, a mode of representation deviating from this format had to be avoided. Thus, the creation of fish failed to be depicted in its manifest form. Instead, all that the artist could manage to do by way of compromise, without breaching the rule, was to let the presence of fish be implied by the modest image of water under the ground.72 That fewer birds and animals are depicted in Junius than in the other two works may be subject to understanding by appealing to the smaller space available to Junius on account of the particular format adopted for illustration, namely, a frieze. More specifically, while groups of birds and animals are distributed densely throughout the available panel at Salerno and San Marco, they are contained tightly, one instance each, within the single strip in Junius. Thus, a frieze cannot afford to accommodate a flock of birds and a herd of beasts such as those allowed to occur freely at Salerno and San Marco. A one–dimensional arrangement is absolutely incompatible with hosting a crowd of birds and animals in a single panel. It should be emphasized at this point that at stake here is the structural organization of a frieze, rather than simply the physical lack of space for representing a multitude of animals. As it is, the lower dome of P7.1 has access to a relatively large space for representation, as its upper part is left blank. Without the requirement of one–dimensionality, the empty space available could readily have been filled with numerous figures of birds and animals.73 72 In view of the artist’s well-motivated decision not to represent fish, Gollancz’s (1927: xl) remark seems to be on the right track: “Possibly the artist had intended to indicate the water sundered from the earth and the creation of fish in the space between the two pictures where faint rippling lines may be noted.” 73 This restriction only accounts for the single procession of living creatures. It remains unclear why a limited number became the minimal number—one—for each species.

4.2 A diachronic–comparative perspective 

 117

Finally, the depiction of plants in P7.1 at all and their multiple appearance at that need explanation. At first glance, the occurrence of vegetation seems contradictory to the absence of heavenly bodies, because neither is a creation of the fifth or sixth day. In fact, neither is represented in the corresponding pictures at Salerno and San Marco. When one invokes the reconceptualization and redesigning that would have been demonstrably in force as argued above, one may gain insight into the apparent incongruity in representation in P7.1. In recomposing and reformatting model pictures reflected in those of Salerno and San Marco, the illustrator had to compress the Creator’s works of the fifth and sixth days into a single frieze under the arch, in conformity with the convention of architectural framing. Moreover, also as a result of verticalization, God had to be represented in the upper sphere, far removed from the scene of creation on earth. These two artistic motivations would have converged to lead the artist to reconceptualize the whole representation as a scene of God beholding and blessing what he had created on earth. Thus, rather than recording exclusively the two days’ work, the P7.1 picture recapitulates all of the achievements—including the creation of vegetation—made so far, up to the creation of animals, on earth. Another noteworthy feature of the plants in P7.1 is their highly stylized form (cf. Blum 1976: 220), to the point that they come close to being perceivable as geometric figures, in marked contrast to the plants in P7.2 that are represented more naturally. Of particular interest is the plant to the left of the bird standing at the middle of the frieze: it is symmetrical and attached with a cross at the center. Further, the plants are evenly distributed with a regular interval in spacing, and two of the three intervening spaces thus available are filled with a bird (center) and an animal (right), with the left one empty. Such a regular alternate pattern of a plant and a nonvegetative figure (including a zero one) seems too striking to be regarded as accidental.74 As argued above, I have adduced several probable models for the arch-based framework of P6 and P7. The canon tables and the ruler or evangelist portraits are of central interest. While these multiple sources hardly allow for individual identification or relative weighing in terms of their likelihood for having served as actual models for the creation images in Junius, a more precise identification and evaluation seem feasible in the case of P7.1. At issue in particular are the two distinguishing features: the geometric composition of the plants, and their regular alternation with nonvegetative figures in friezing. Both of these design properties are derived most likely from the format of a canon table. The plants (notice that there are usually more than one instance depicted), birds, and animals are often found on the rim of an outer arch or at the peak and/or trough of the rim of an inner arch, as at the Harley Golden Gospels (London, British 74 One might be tempted to speculate that the zero form stands for fish under the water. The strict adherence to a frieze-based representation disallows fish to be depicted in the water, as argued above in the text. Yet the inclusion of fish in the frieze would have been rejected as unnatural. Thus, by way of compromise the artist may have alluded to fish by a blank space in the frieze, the only format available for representation.

118 

 4 The Creation: P6/P7

Library, Harley MS 2788), fols. 6v–11v and Vivian 326r–327v (Pl. 29).75 The Vivian canon tables are of outstanding significance, in which the plants, executed in their excessively stylized form, are set at the troughs of the smaller arches, with their bodies appearing above the height of the peak of the arches. These plants are thus arranged at regular intervals, exactly the type of configuration that shapes the frieze in Junius, in which the original rims of the inner arches under the decorative plants are simplified as plain ground.76 In this connection, the specific plant in Junius standing next to the bird, referred to above, deserves particular attention. Its symmetrical form closely resembles the vegetative figures that are found at the center on the rim of the outer arch at Vivian 326r (the central part), 327r (with a cross), and 327v. In light of these formal and substantive parallels ascertained between Junius and Vivian, one may be led to infer with reasonable confidence that the artist responsible for the innovative representation of the creation pictures in Junius was inspired critically by the composition of the canon tables in Vivian (and/or others closely akin to them). At this point, it may be appropriate to further substantiate the Touronian foundations of P7.1 through a comparison with Utrecht 82v (Pl. 81). A visualization of Psalm 148, this Utrecht picture provides a panoramic view of heaven and earth and all the creatures that God established. Given that P7.1, too, is intended to be a representation of God’s creation (as argued above), it will be illuminating to compare the two illustrations and thereby bring into focus the distinguishing features of the Junius picture that are foreign to its Utrecht counterpart. Despite the largely common interests in subject matters for visualization (humans are disregarded here, as they fall outside the scope of P7.1), the two works markedly diverge in manners of representation. Overall, Junius constitutes a radically simplified and reduced version of God’s creation, in contrast to the encyclopedic nature of Utrecht, as follows. First, no angelic beings are depicted in Junius. Particularly noteworthy is the God figure not being flanked by angels, in defiance of the conventional configuration to the contrary. The absence of angels in Junius derives from the corresponding nonexistence in the model, the Salerno plaque (or other works embodying the same tradition). Second, there are no celestial bodies—sun, moon, stars, and winds—in sight, although the circumference of the firmament is appreciably drawn. The exclusion of heavenly bodies would have been motivated by the Junius artist’s strict concern with creations on earth in P7.1, in contrast with P6 and P7.2, where God’s work of creation also has bearing on heaven. Third, earthly lives are treated divergently between the two pictures under examination. While Utrecht provides a natural landscape full of living creatures in a seemingly two–dimensional format most fitting

75 For the long tradition of devising canon tables in the Late Antiquity and the Early Middle Ages, see Nordenfalk (1938). 76 As far as the representation of the ground and the water beneath is concerned, the Harley Golden Gospels (fols. 7v–10r, 11r–11v) provides appreciable parallels to Junius: above the smaller arches, the smooth ground consisting of several layers is laid out in a way reminiscent of the water in Junius.

4.2 A diachronic–comparative perspective 

 119

the scene,77 Junius follows a frieze pattern in such a consistent way that the creatures selected for illustration are aligned on the same baseline. As a consequence, a bird has to be shown resting on the ground in Junius, whereas some birds are naturally depicted flying in the air or perching on the tree in Utrecht; and, while the one–dimensional format excludes fish from representation in Junius as discussed above, sea creatures are fully eligible for illustration in Utrecht in the absence of comparable restrictions on composition. Fourth, the plants in P7.1 look rather like abstract geometrical objects and stand on their own in isolation from others. By contrast, the trees in Utrecht are of a natural form and integrated into a whole living sphere full of birds and animals. The apparent artificiality and lack of organic naturalness characteristic of the Junius picture should not be confused with the artist’s technical ineptness; rather, it is a consequence of his coherent conformity with the particular format that he correctly appreciated to be traditionally inherent in the model framework of his choice: the canon tables from which the artist drew inspiration allow for a frieze on the rim of the outer and/or inner arches, in which the stylized, more or less symmetrically shaped, vegetative figures are placed for decoration at regular intervals, often accompanied by birds and/or animals. The diametrical opposition between P7.1 and Utrecht 82v in representing essentially the same subject matter has significant implications for assessing possible influences of the Utrecht Psalter on Junius 11. Despite the apparently ready-made images, the Junius artist (or his antecedents) obviously did not venture to use any of the materials thus available. It is hardly conceivable in this regard that the Junius artist was simply denied access to the Utrecht pictures. On the contrary, there are appreciable superficial resemblances between the two works, such as the double processions of angels in P3.1 (section 2.2). In this light, one may suspect that the total absence of similarity between P7.1 and Utrecht 82v was due to a deliberate avoidance on the part of the Junius artist. Constrained by consequences of his choice of the canon table format, he found every reason not to adopt any of the Utrecht images for his use. In this way, the Utrecht Psalter would have served as an inventory of images that the Junius illuminator laid his hand on, insofar as they happened to suit the artistic intentions and preferences that he had independently cherished. The Utrecht Psalter in itself hardly played a critically formative role in leading the Junius artist in the course of his artistic activities.78 77 To be fully discussed in section 5.2 below, the actual format underlying the Utrecht picture is a multiple of one–dimensional rows, rather than a two–dimensional plane. For present purposes, the impressionistic loose labeling is enough to highlight the difference between the strict linearity (P7.1) and the multiplicity of friezes (Utrecht 82v). 78 Section 5.2 will examine another case in which Utrecht 82v failed to influence a comparable representation of God beholding his creations, namely, P11. In brief, Utrecht 82v would have been rejected as unsuitable for P11 because of its excessively schematic character, whereas the same picture would have been rated as too naturalistic to meet the convention at hand in the case of P7.1. Thus, on the basis of the divergent criteria properly motivated in each context, Utrecht 82v was denied contribution to the composition of the two closely related pictures in Junius, P7.1 and P11.

120 

 4 The Creation: P6/P7

Before bringing this section to an end, Blum (1976) deserves detailed evaluation. A radical appraisal of P6 and P7 together with the accompanying inscriptions, Blum’s view is in many respects diametrically opposed to mine as submitted above. While being a convoluted complex of hypotheses intertwined with conjectures and speculations scattered throughout, Blum’s thesis is in its essence composed of two major claims. Although presented as a unitary set of mutually supporting, inseparable statements, the two hypotheses are logically and empirically independent of each other, and accordingly require separate assessments. First, the creation cycle in Junius belongs to the Octateuch tradition. Second, the inscription at page 6 of the Junius manuscript should be read as “Before he separated water and earth.” These core claims are supplemented, reinforced, and elaborated with ancillary interpretations that are devised to buttress Blum’s whole interpretive framework. Blum (1976: 214) starts by expressing her wholesale commitment to Pächt’s (1943: 62) observation: P6/P7 and the creation pictures of the Egerton Genesis (London, British Library, Egerton MS 1894; Joslin/Watson 2001) at fols. 1r–1v are derived from a common source. Pächt’s view, however, appears tenuous, as it is based solely on the common property of the Creator sitting on an arc high above the scene of creation, in defiance of numerous dissimilarities (for details of these features characteristic of Junius, see above): (i) the arc as the firmament (Junius) versus the rainbow (Egerton);79 (ii) the presence (Junius) versus absence (Egerton) of angelic beings (compare Blum’s inconclusive remarks on this divergence; Blum 1976: 224n30); (iii) the sitting on the inner arc (Junius) versus the outer arc (Egerton); (iv) complementariy in representation: the separation of light and darkness is represented only in Junius, whereas the water above and below the firmament is treated only in Egerton; (v) the absence (Junius) versus presence (Egerton) of the outer frames (or alternatively formulated, the circular (Junius) versus square (Egerton) frames;80 (vi) the selective (Junius) versus additive (or cumulative; Egerton) mode of representation (by the latter mode, the works of earlier days are inherited in representation for later days); (vii) representational economy (related to selectivism as [vi] above, Junius), as with the absence (Junius) versus presence (Egerton) of clouds and rainbows, which are irrelevant to the subject matters; (viii) the semicircular (Junius) versus fully circular (Egerton) form of the earth.81 Strictly speaking, Pächt’s view has no intrinsic bearing on Blum’s thesis, which can be sustained regardless of the subsumability of the creation cycles in Junius and Egerton under the same tradition. Perhaps, in Blum’s view, the alleged existence of a parallel case of the Octateuch tradi79 The firmament is represented within the globe below the rainbow on the second day (fol. 1r, upper right) and afterwards in Egerton. As remarked in Joslin/Watson (2001: 30), the image of the Creator sitting on a rainbow is unique to Egerton. In view of these discrepancies, the apparent similarity of being seated on an arc does not seem to be specific enough to connect Junius and Egerton. 80 This contrast may be derived from feature (vii) below. The criterion of representational economy may dismiss separate use of an outer frame, as the two existing arcs are sufficient to enclose all figures. 81 Perhaps this difference may be reducible to the representational economy (vii) above, which motivates the exclusion of the lower circumference that is irrelevant to the landscape of creation.

4.2 A diachronic–comparative perspective 

 121

tion in England—the Egerton Genesis—would enhance the plausibility of her identifying the Junius creation cycle as an inheritance of that tradition: there are putatively at least two exemplars of the eastern tradition attested in England. Building on this questionable view of Pächt, which Blum (1976: 214, 219; cf. 224n30) takes to be an indisputable truth, she identifies the common English source in question as belonging to the Octateuch tradition (Blum 1976: 215), one of her two major claims. Blum (1976: 217–219) then attempts to substantiate this identification in regard to a few figural traits of P6/P7.82 Brought to attention by Blum are the following rather isolated, abstract geometrical features lacking specificity and systematicity: (i) the irregular horizontal lines signifying water (P6) or ground (P7); (ii) the “spidery” stars set in the sky (P7.1). Feature (i), the irregular ground- or waterline, seems ubiquitous, not only in the Octateuchs, but also in the Cotton Genesis family and elsewhere on Carolingian soil, as in Vivian 3v (Pl. 26), the Ebbo Gospels 18v, and the Codex Aureus 16r. This ubiquity may call into question the alleged genealogical relation between P6/P7 and the Octateuchs that Blum would like to posit in defiance of analogous correspondences attested beyond. Of further interest in this connection are the Codex Aureus 16r and the Ebbo Gospels 18v, 60v, and 90v (the ground under the evangelists): not only parallel to P7 in the contour of the ground, they also exhibit plants sprouting from it, which are strikingly similar in appearance to those found growing in P7.1, as discussed above. In this light, it may be more legitimate and illuminating to identify source materials within the confines of Carolingian manuscripts in line with the other pages examined so far than to go so far as to the Greek Octateuch tradition. With regard to feature (ii), the “spidery” stars appear to be unique to Junius. The corresponding objects are shaped more carefully and exquisitely in the Octateuch and the Cotton Genesis tradition. Moreover, the randomly scattered stars in Junius contrast with the more or less evenly distributed ones elsewhere. Thus, the appeal to such figural and compositional features for justification of either source lacks immediate relevance. Correspondingly, Junius is distinguished from exemplars of both traditions by the absence of the sun and the moon in their explicit form. In short, as with the ubiquitous groundline but for the opposite reason of uniqueness, the heavenly bodies cannot afford to shed light on the derivational relation between the creation pictures in Junius and their corresponding cycles in other works.83 82 For further criticism of Blum’s view, supplementary to what follows below, see Broderick (1978: 127). 83 In addition, in support of her claim Blum (1976: 216) compares the covering angel in P6 with the personification of night in the Octateuchs, particularly the figure’s gesture of raising the left hand close to the face depicted in Vaticanus Graecus 747 (Vatican, Biblioteca Apostolica, Vat. gr. 747), fol. 15r (Weitzmann/Bernabò 1999: pl. 21; see also Henderson 1975: 142 for a similar observation on another Octateuch, Seraglio codex, Istanbul, Topkai Sarayi Library, codex G. I. 8, 26v; Weitzmann/Bernabò 1999: pl. 18). The alleged similarity, if it is allowed to be so called, is insubstantial and superficial, limited as it is simply to the proximity of hand to face, not involving particular individual features and their configurations. After all, the Octateuchs exhibit no remarkable resemblances to the Junius creation pictures in specific points of detail, as Broderick (1978: 139–140) has correctly pointed out.

122 

 4 The Creation: P6/P7

In conclusion, in the absence of compelling evidence for the Octateuch influence and conversely in the presence of demonstrable grounds for positing the Cotton Genesis and Touronian Bibles as models, as substantiated in depth above, it may stand to reason to generalize that the creation pictures in Junius would primarily have been inspired by the Cotton Genesis tradition in general, and its manifestations in Carolingian manuscripts in particular, notably those from Tours.84 With respect to the Octateuch background that she postulates behind the production of P6 and P7, Blum makes the following inference, a further assumption to her first claim. Given the allegedly Octateuch underpinnings of the creation cycle realized in Junius, the scribe would have intended to follow the Octateuch exemplar with maximum fidelity, as attested in the extant group of six manuscripts (Weitzmann/Bernabò 1999) in designing his work. At this point, Blum makes the second major claim by proposing an alternative emendation to the inscription on page 6 (Blum 1976: 215): instead of the usual reading “[Hēr h]e” (here He . . .), Blum (1976: 215) posits “[Ǣr h]e” (before He . . .). Since her proposal has no intrinsic advantage on its own,85 its eventual legitimacy and acceptability must depend on how explanatory and feasible it proves to be in light of empirical consequences that it entails, to be argued in the following. Drawing on the exemplary status of the Octateuch source as a model for the Junius artist as hypothesized above, Blum infers that the proposed reading of the first inscription expresses the scribe’s original intention: Page 6 was originally meant to represent the world before creation, comparable to the first picture in Egerton and the Octateuchs. Furthermore, interpreting the second inscription at page 7 “Here He divided day from night” as an instruction to provide an illustration for the first day (day = light, night = darkness), Blum (1976: 224n24) reconstructs the originally planned distribution of the creation pictures (including the creation of Adam) at pages 6 through 14 (three leaves—pp.  9–14—subsequently lost), on the basis of the existing series recorded in the Octateuchs. In this connection, Blum claims that the Octateuch tradition suits the available space in the original manuscript (ranging from pages 6 to 14) better than the Cotton Genesis practice (as exemplified at San Marco) in terms of the number of panels required. This argument must be dismissed as inconclusive. Strangely enough, Blum makes no mention of Salerno and other exemplars of the Cotton Genesis recension, as treated above, which contain a smaller number of pictures, more or less indistinguishable from the Octateuchs in the number of pictures used. The purported closer correspondence to the pictorial space according to Blum’s speculation, therefore, cannot be adduced as a corroboration of the Octateuch source. Moreover, since the Cotton Genesis tradition demonstrably underlies P6 and P7, as examined above, and since it also contains a picture of the universe before the separation of light from darkness, Blum’s 84 Blum (1976: 217) herself acknowledges the Junius artist’s indebtedness to the Cotton Genesis tradition with regard to the angelic figures of light and darkness. 85 Rather, it is suspected as less plausible than the standard view, given that there are a number of parallel inscriptions in Junius that begin with Hēr, but none starting with Ǣr (compare Lucas 1980: 209n58).

4.2 A diachronic–comparative perspective 

 123

reinterpretation does not necessitate the postulation of the Octateuch tradition behind Junius, as it is as compatible with its Cotton counterpart (cf. Broderick 1978: 127). The existing illustrations in Junius P6 and P7 are evidently at odds with Blum’s claim, as they cover all works before the creation of Adam. Blum is thus compelled to explain the discrepancy between the alleged original plan and the actual execution. In Blum’s view, the lack of space resulting from loss of the three leaves would have made it impossible for the artist to put into action the instruction he received. In order to make most of the reduced space still available (pages 6 and 7), he had to condense the creation pictures into the three scenes, as they stand in the extant manuscript (Blum 1976: 215, 224n24, n25).86 Thus, without bringing about any explanatory accounts at the cost of incurring unnecessary conceptual complications lacking empirical basis, Blum’s scenario—to the exclusion of other possibilities as mentioned by Broderick (1978: 122– 123)—reverts all the way to the starting point that one would otherwise have already departed from for substantive exploration: the motivations and mechanisms for the ­artist’s innovative reorganization of traditional materials, as investigated above. In other words, Blum’s emendation of the first inscription simply creates quasi problems otherwise uncalled for and leads nowhere toward their explanation. Blum’s reinterpretation of the first inscription, taken on its own, has implications that are devastating for her own framework of explanation. According to a natural interpretation, “before the separation of water and earth” (occurring on the third day of creation) would mean all the pictures up to and including the second day, rather than exclusively the initial illustration of the cycle (for a similar criticism, see Lucas 1980: 209n58). Indeed, the second inscription on page 7 may serve to correctly exclude such an extensive range of referents, but it would hardly be appropriate to use the expression “before the separation of water and earth” for the purpose of making an exclusive reference to the world before creation, contrary to Blum’s wishful reading, which would make sense only in anticipation of the inscription to follow on the next page. To make such a precise identification on its own, one would best invoke the separation of light (day) from darkness (night)—the work of the very first day—as an optimal terminus ante quem. As it is, the reconstructed instruction falls short of the maxim of quantity—“be informative enough as is necessary”—and the one of manner—“be clear in presentation”—for successful communication (cf. Grice 1975): less informative and clear than is required, it cannot function as an appropriate instruction. Unless specifically motivated otherwise, such an inappropriate command would unlikely be formulated on purpose. This line of pragmatics-based reasoning, then, would cast doubt on the plausibility of Blum’s interpretation on the whole: one is hardly expected to articulate one’s request deliberately in a vague and imprecise fashion. On the contrary, if the scribe’s inten-

86 Such a reasoning presupposes that the three leaves were lost after the scribe gave his instruction to the artist and before the latter began his work of illumination. This would be an interesting empirical implication worthy of verification.

124 

 4 The Creation: P6/P7

tion had actually been as Blum conjectures, he would most likely have formulated his instruction as “before He separated light (or day) and darkness (or night).” Alternatively, and less efficiently, he may have left the pictorial space on page 6 without a note, and been satisfied instead with letting a single instruction on the following page convey his intention, “Here He separated day and night.” Thus, Blum’s second claim would have to entail that the scribe was engaged in an extremely unlikely speech act of making a request incompetently. To make the matter still worse, “Here He separated day and night” itself is ambiguous, as it is open to two competing interpretations. First, as read by Blum (1976), the statement may refer to the separation of light (called Day) from darkness (called Night) on the first day of creation. Alternatively, however, the sentence may be construed as relating to the fourth day on which the sun and the moon were set in the sky to rule the day and the night, respectively. In this way, the second inscription proves to be no less vague in its content than the first, insofar as both inscriptions are understood to be the scribe’s instructions to the illuminator, following Blum’s proposal. One must accordingly be led to assume that the scribe was scarcely competent as a linguistic communicator, an extremely unnatural inference. By contrast, on the reading proposed above—that the inscriptions were intended as explanatory notes supplied to the pictures—the inherent ambiguity in reference would cease to be a problem, simply because the corresponding images help determine the intended meanings encoded in them. The major thrust of the above exercise in verbal interpretation with regard to the two inscriptions accompanying the P6 and P7 illustrations should by now be self-evident: as proper speech acts, they must not be taken to be directives; rather, they are designed to be assertives or descriptive statements.87 The foregoing criticism has been presented, for the sake of argument, on the basis of Blum’s assumption that the Junius artist followed the Octateuch tradition as a model for the creation pictures. The proposed emendation does not support the postulation of the Octateuch behind Junius, as it would entail an infelicitous speech act on the part of the scribe that would have obstructed his intention from being carried out properly. Paradoxically, Blum’s reinterpretation would make better sense under the hypothesis, contrary to Blum, that the scribe designed, from the start, the arrangement of creation pictures such as were actually implemented in the manuscript. Specifically, the P6 picture covers the first two days of creation, that is, the works that had been completed before the separation of water and land that was to take place on the third day. The inscription, then, implies correctly that the following scenes on page 7 represent a series of events that follow the reference act of creation at issue. The absence of a specific illustration for the separation of water and land is readily intelligible in light of the verbal information conveyed by the inscription that would rather make its visualization superfluous. Thus, Blum’s second claim, the radical rereading of the inscription, does

87 On these pragmatic notions in Speech Act Theory, see Austin (1962) and subsequent works in the field.

4.2 A diachronic–comparative perspective 

 125

not necessitate the Octateuch tradition; on the contrary, it would exclude this source as being unlikely in the extreme. In sum, ingenuous as it may appear at first glance, Blum’s thesis proves to be a bewildering array of casual remarks, nonexplanatory stipulations, far-fetched conjectures, and insubstantial speculations, loosely connected by convoluted reasoning and stretched comparison; and what is still worse, her double core claims are hard to reconcile with each other. In conclusion, one may well dismiss the Octateuch tradition as unlikely to have played a single determining role in the construction of P6/P7, and reject Blum’s alternative reading of the first inscription as incapable of being more explanatory than the conventional interpretation.

5 The Creation of Eve: P9 5.1 A synchronic–structural perspective Concerned with the creation of Eve as a whole, P9 (Pl. 6) comprises three scenes. Each picture in turn is supplied with an inscription above describing the scene. At the lower right, God is bending over the sleeping Adam and removing a rib from him: “Hēr drihten gewearp sclēp on adam and genam him ān rib of þā sīdan. And gescōp his wīf of þām ribbe” (Here the Lord cast a sleep on Adam and took a rib from his side and created his wife from the rib; Gollancz 1927: xl). To its upper left in the left half of the middle zone, God is lifting up the created Eve: “Hēr drihten gescōp adames wīf euam” (Here the Lord created Adam’s wife Eve; Gollancz 1927: xl). To its upper right in the top area stands a host of angels in heaven, with Michael—as identified in the annotation above his head—in the center at the open gate. On a ladder bridging heaven and paradise an angel is descending: “Hēr godes englas āstīgan of heouenan into paradisum” (Here God’s angels descend from heaven into paradise; Gollancz 1927: xl). Obviously, the first two scenes are temporally ordered, while the last picture may be broadly conceptualized as a background to God’s acts of creation on earth. The upper right scene implies that God had descended by ladder from heaven, followed by an angel who is shown being on his way. In the meantime, Michael, God’s deputy, temporarily came to occupy the central position at the door. Accordingly, it may be inferred that earlier God had been located at the center in heaven, attended by five angels on either side. At the moment, the host of angels beholds the creation of Eve from above, as illustrated. Thus, the upper right image in heaven frames in temporal terms the lower right and upper left scenes in paradise. The three constituent pictures are distributed on the page conjointly without specific boundaries between them. As a consequence, the tree located at the lower left periphery is ambiguous as to its localization: Is it part of the first scene or the second? Or, does it belong to both or neither, intended to function as a boundary? Had the tree located between the two figures of God, it clearly would serve as a boundary between the two scenes. However, connecting this tree with the slanted ladder results in a diagonal line that neatly divides the two scenes in paradise. In this light, one may legitimately view the tree as part of the boundary marker. A number of pictures are contrasted with and analogous to the complex of images in P9 in varying manners and to varying extents. By far the most outstanding in contrasts and correspondences is no doubt P20 (Pl. 12; section 7.1). In conceptual terms, the two pages constitute polar opposites. While P9 is primarily concerned with the creation of Eve, P20 is devoted to the torture of the fallen angels in hell. Given that Adam and Eve were created by God to occupy the kingdom of heaven as a substitute for the fallen angels, which enraged and was lamented by Lucifer, particularly in Genesis B 365a–368a and 395a–397a, the human couple and the devils are absolutely antagonistic in terms https://doi.org/10.1515/9783110788068-005

128 

 5 The Creation of Eve: P9

of God’s love and hatred, respectively. Corresponding to this fundamental conceptual opposition, P9 and P20 are naturally differentiated in almost all conceivable ways, as substantiated in the following paragraphs. At the same time, however, both are analogously composed in their structural foundations and exhibit corresponding formal properties. P20 is comparably tripartite in its pictorial representation, one scene in hell and two in paradise, corresponding to one scene in heaven and two in paradise in P9; and between the two worlds, hell and paradise, an additional bordering space is provided for the interface in P20, corresponding to the connecting ladder between heaven and paradise in P9. On these common bases, however, P20 diverges from P9 along a cluster of parameters, as substantiated in the following. Some of the differences may be derived mechanically from the distinct conceptual bases underlying the two pictures. Thus, paradise occupies the lower (P9) and upper (P20) half of the page. Correspondingly, an angelic being descends from above, whereas a devil moves up to paradise. All the remaining parameters, however, are specified independently of the underlying conceptual grounds and in maximally contrastive manners. For the sake of exposition, the parameters involved may fall into two categories, compositional and figural. The first class concerns a spatial arrangement of chronologically ordered events on the one hand, and a configuration of angels/devils in heaven/hell on the other. As far as the former parameter is concerned, P9 may be characterized by a diagonal right-toleft direction in the paradise scenes, as God’s taking of a rib from Adam at lower right is followed by the lifting of Eve at upper left. By contrast, P20.1 (paradise scenes in the upper half) is oriented in the straight horizontal right-to-left direction: first on the right Adam and Eve together face toward God in their full obedience (though God himself is unrepresented here) and away from the forbidden tree and, by implication, from Satan’s messenger, in defiance; and second on the left side, originally Adam (but here altered as Eve secondarily and by mistake; section 7.1) is tempted in vain by Satan’s agent (for details on this apparently ambivalent picture, see section 7.2 below). Thus, the paradise scenes on the two pages are contrasted by the features of being diagonal (P9) versus strictly horizontal (P20). And yet there is a further dimension to the directionality issue, which involves the direction of movement into paradise. In P9, the passage from heaven to paradise is slanted, as shaped by the ladder that connects the two spheres. In contrast, the corresponding passage from hell to paradise is straight and vertical, as indicated by the posture of the departing devil emissary. Thus, to generalize, the oblique and straight directionalities fundamentally distinguish between P9 and P20 in their pictorial organization. As regards the second compositional parameter of arranging angels and devils in heaven and hell, respectively, P9 and P20 are no less contrastively composed—orderly (P9) and disorderly (P20). In P9, the host of angels are arrayed in heaven in such a regular manner that Michael stands in the center, flanked on both sides by groups of fellow angels facing each other. In P20, with a single exception, none of the devils stands in a normal upright position; they assume various poses, flying rightward or leftward,

5.1 A synchronic–structural perspective 

 129

falling, and prostrate. They are scattered in the enclosure without forming a consistent overall pattern even remotely comparable to that in P9. Figural contrasts and correspondences primarily apply to the representations in paradise. Beginning with an entrance to Eden, at the interface between heaven and paradise, as depicted in P9, a large door is wide open, from which a ladder leads to paradise, on which a heavenly being may ascend or descend upright. By contrast, hell connects with paradise through a narrow opening similar to an escape hatch, which a devil manages to creep through. Thus, the interface between heaven and paradise, including the ways in which one may pass through, is distinguished from that between hell and paradise by a set of opposite values, large versus small gates, open versus closed gates by default, God’s versus Satan’s agents shown crossing the border, and their standing versus lying postures on their way to paradise. When it comes to the landscape of paradise itself, two pairs of figures appear in each illustration of paradise: God and Adam, and God and Eve in P9 on the one hand, Satan’s messenger in the form of a snake and Adam,88 and Adam and Eve in P20 on the other. The left pair on each page confronts each other, while the right counterpart faces the same direction. Furthermore, the right pair makes closer contact with each other than the left one, as embodied by a shorter distance involved between the constituent figures: in P9 (right), God bends over Adam and takes a rib from him, while in P20 (right) Adam and Eve are represented very much overlapped, suggesting their unified mode of existence in terms of their perfect fidelity to God. In P20 (left) Adam, standing on the right, looks down on Satan’s messenger to the left, much as in P9 (left) God, standing on the right at a higher level than Eve on the left, raises her up. Overall, precisely as God, represented twice on the inner side of the two constituent pictures, dominates both scenes in P9 as agent of creation, Adam, comparably appearing on both sides of the central tree that serves as a dividing line in the upper half of P20, controls the two events depicted on the right and the left through his strict observance of God’s commandment. P17 (Pl. 11) may be mentioned as an example exhibiting contrast and parallelism with P9 on a less conspicuous scale. Comparably divided into upper and lower halves for illustration, the picture represents God, accompanied with seraphim on both sides, seated on a throne in heaven, while down below in hell a crowd of devils throngs in chaos with Satan at the bottom. In contrast to the interface between heaven and paradise in P9, the two spheres are completely separated from each other in P17, without a connecting passage between—no door, no gate, and not even an escape hatch. While the lower sphere in P9—paradise—is depicted as an open field with a depth and an extension, hell constitutes a tightly-enclosed small confine in the corresponding zone of P17. Paradise is readily accessible to heavenly beings, whereas hell defies access categorically. Thus, the circular smooth borders between heaven and paradise on the

88 For a precise identification of this figure as Adam in distinction from Eve as would appear apparent, see section 7.1 below

130 

 5 The Creation of Eve: P9

one hand (P9), and the walled enclosure attached with crenellations on the other (P17), embody the fundamental difference between paradise and hell in their demarcation from heaven. Correspondingly, God is engaged in acts of creation on earth, while he is simply seated on the throne totally aloof from the scene in hell. Thus, while heaven and paradise are shown closely interrelated, heaven and hell lack mutual accessibility, fully autonomous on their own. As observed at the beginning of this section, Michael’s presence at the gate would imply that God had earlier stood at this place before descending the ladder. Moreover, the shape of the open gate is reminiscent of a mandorla surrounding the Creator. In this light, the image of angelic beings arrayed in heaven and headed by Michael may invoke by association the motif of God enthroned. Thus, P9 and P17 are linked on the basis of the common heavenly scene, against the background of which paradise (P9) and hell (P17) are represented by contrast as the foregrounded sites for the thematic events to be illustrated. The representation of Michael flanked by accompanying fellow angels in heaven may evoke another image by association and opposition: Lucifer receiving palms from (or giving them to) the arrays of angels on both sides (P3.2a; section 3.1).89 This association will further extend to the fall of Lucifer, P3.3 and P16, and finally to his torment in hell, as represented in P17. In this way, P9 articulates with P17 through multiple channels of correspondence and contrast.

5.2 A diachronic–comparative perspective Distinctively characteristic of P9, the tripartite, triangular arrangement of individual scenes in a single picture finds a close structural parallel in the left half of the first register of Vivian 10v1 (Pl. 27): the creation of Adam at the left, the removing of a rib from Adam at the lower right, and the angel praying with outstretched arms and observing the creation of Adam at the apex. It should be noted that these three scenes may be conceptualized as an integrated whole in Vivian, by virtue of a tree that serves to delineate the left half in question as a whole from the succeeding scene to the right. With an aim to representing the creation of Eve, the Junius artist would have come to appreciate the tripartite configuration in Vivian 10v1 (left) as an underlying scheme for his work, to be detailed in due course. There are, however, five outstanding disparities between P9 and its alleged model, which would make inviable a straightforward derivation of the former from the latter— something closest to being a simple copy. First is the thematic difference at large: while

89 With an alternative but less plausible (and eventually rejected) interpretation that the central figure is Michael (section 3.1), the conceptual connection between P9 and P3.2a would be maximally strengthened. The upper panel of P9 would accordingly be characterized as an elaborate variant of P3.2a.

5.2 A diachronic–comparative perspective 

 131

the creation of Eve is at issue in Junius 11, that of Adam, followed by a beginning of Eve’s creation, is illustrated in Vivian. Second, corresponding to the replacement of Adam with Eve, God’s position relative to the human is reversed: while God appears to the right of Eve in P9, he occupies the opposite position relative to Adam in Vivian 10v1. Third, the paradise scenes proceed from right to left in P9, but conversely from left to right in Vivian 10v1. Fourth, the triangular layout of the three constituent scenes in Vivian is transformed in P9: the two diagonally arranged paradise scenes, as opposed to the strict horizontal alignment in Vivian 10v1 and P20 (the conceptual and compositional opposite of P9; section 5.1), shift the center of the angelic host to the right, thereby resulting in an overall placement in a distorted triangular form. Fifth, a host of angels is depicted in P9, whereas only a single angel is on the scene in Vivian 10v1. The identification of the left half field of the top register in Vivian 10v as the model for P9 is thus challenged to provide explanatory accounts of these discrepancies in the course of deriving P9 from the alleged source. Since a common starting point for the narrative of the creation of Eve is constituted by the scene of God taking a rib from Adam as represented in the Cotton Genesis cycle (Kessler 1977: 17)—Bamberg (Pl. 23), Grandval (Pl. 24), Vivian (Pl. 27), San Paolo (Pl. 34), San Marco (Pl.  43, left), and Millstatt (Pl.  62), the Junius artist would naturally have chosen the event as an initial phase of the story. The next stage concerns the shaping of Eve, as depicted in San Paolo and San Marco (Pl. 43, right). This scene, however, is absent in Vivian, directly followed instead by the introduction of Eve to Adam, as realized on the right side of the first register. Fully familiar with the shaping phase coming in between despite its absence in Vivian, the illuminator reinterpreted the animation of Adam to the left in the Vivian model as the forming of Eve, by drawing on comparable illustrations as in San Paolo 8v2 (left, Pl. 34). In this way, the radical redefinition as the creation of Eve was accomplished on the model of the creation of Adam.90 Reworking thus the paired figures of God/Adam in Vivian 10v1 as Eve/God in P9, motivated by the reconceptualization discussed above, necessitates reversal of the leftto-right to the right-to-left narrative development, due to the temporal order imposed whereby God takes a rib from Adam (at right) before he creates Eve (at left). It may be pointed out that the reversed order thus required proves to be compatible with the existing pattern of sequencing that is independently implemented elsewhere in Junius 11, namely, P3.1 (section 3.1), P41.1 (section 10.1), and most significantly P20 (section  7.1), the counterpart that P9 is polarized with for maximal contrast. This reordering in turn induces inversion of the pairing God/Eve—the immediate result of the reinterpretation—to Eve/God, in accordance with a pivotal role of God as agent

90 This would suggest that the Junius artist had access to, as well as familiarity with, a range of works belonging to the Touronian Bibles/the Cotton Genesis family members, rather than having at his disposal an isolated knowledge of a single specimen of the Touronian Bibles, namely, the Vivian Bible. Implications of this finding will be explored in Chapter 12.

132 

 5 The Creation of Eve: P9

of creation in integrating the two separate acts as constitutive of the creation of Eve. The representation of double-agency through a largely mirror-image duplication of a central figure is independently executed in P44 (section 10.1), which may be adduced as evidence of the artist’s preference for this pattern of composition. Whether intentional or coincidental, the resulting configuration constitutes a reversal along the horizontal axis of the scene represented at San Marco (Pl. 43). Thus, the second and third deviations from the Vivian model, that is, the reversal of God’s position relative to Adam or Eve, and the inversion from left–right to right–left development, have been subject to principled explanations. The fourth peculiarity characteristic of P9, namely, the slanted, upward movement from the first scene to the second in P9 instead of the strictly horizontal advancement in Vivian (and elsewhere in the Touronian Bibles) may be ascribed to a geometrical balance. More specifically, the diagonal axis shaped by the ladder would have motivated a relative positioning of the two scenes in paradise so as to achieve a largely symmetrical composition. That is, aligning the two events with a baseline perpendicular to the axis running from the gate of heaven to the ground of paradise, gives rise to the oblique narrative transition from lower right to upper left that is implemented in P9. It may also be of interest to note that the preference for a well-balanced allocation of the two scenes to the left and right of the slanted axis referred to above would have converged to place God beside the ladder by inverting the original configuration of God/Adam to Eve/God via God/Eve. In this connection, one has yet to address Eve’s peculiar posture that is opposed to Adam’s in the original: Eve is turning back and looking up to God, with both hands raised, whereas Adam confronts God at the same height, with his arms stretched downward in the Vivian model. In general terms, the artist would have striven for a polarization of contrast with Adam in appearances, in harmony with the radical redefinition of the figure as Eve: looking backward (Eve, P9) versus looking forward (Adam, Vivian 10v1) to God; standing on a different level (Eve, P9) versus on the same level (Adam, Vivian 10v1) with God; and the hands upward (Eve, P9) versus downward (Adam, Vivian 10v1). More particularly than the formal maximization of figural contrasts, Eve’s specific posture in P9 may be ascribed to the artist’s preference for her analogous gestures as embodied in other pictures in Junius, such as P10, P11, and P13 (section 6.1). So far, it was taken for granted that the artist adapted the left side of the first register in Vivian 10v1 to illustrate the creation of Eve. At this point, it must be explained why the illustrator selected for his reworking this complex of images in Vivian 10v1, rather than simply the single image of God Taking a Rib from Adam. Of central importance in this regard is the appearance of an angel in the Vivian model. In contrast to the Bible, the Genesis A text specifically refers to angels at 185b, on page 9, just before the picture at issue (P9) is inserted: “hēo wǣron englum gelīce” (they [Adam and Eve] were like angels). Immediately below the line that ends with the verse in question is visualized a host of angels in heaven on page 9 (Karkov 2001: 57). It must be called to mind in this

5.2 A diachronic–comparative perspective 

 133

regard that P9 is concerned with the creation of Eve. The presence of angels at the scene of creation applies to the creation of Adam in Vivian and San Marco, not to that of Eve. In contrast, the Millstatt Genesis (fol. 9v, Pl. 62) represents an angel at the site of God removing a rib from Adam. Yet in light of the location of the angels in heaven and the direction in which the ladder extends from heaven to earth in P9, the heavenly scene seems to be primarily associated with the raising of Eve, rather than the taking of Adam’s rib. Accordingly, the close similarity to Millstatt should be regarded as more apparent than real. Although the representation of an angel in Vivian, needless to say, is motivated on other traditional—biblical and extra-biblical, i.e., Jewish—grounds (cf. Kessler 1977: 28–35), it seems reasonable to infer that the textual reference to angels in the Old English/Old Saxon Genesis, more cogently than anything else, would have driven the artist to adapt and reuse the pictorial portion under consideration for his interests, particularly when one may be reminded that the angel is on top of the pictorial complex and located closest to and directly below the corresponding text. Of no less significance is the same posture taken—an orant gesture—of the Vivian angel as Michael’s in P9. This gestural identity may enhance the plausibility as proposed here of deriving the host of angels from the single figure in Vivian 10v1 by amplification. The Vivian angel provides a core element for elaboration, which keeps its original pose intact despite the radical expansion of the company. Furthermore, given that the angel observes the creation of Adam—the left scene, rather than the right—the artist would have been prompted to treat the three images as an integral set, the conceptualization that also would have been facilitated by the presence of the tree as a divider in the center, as discussed above. In other words, segmenting a single image or two from the tripartite pictorial complex in question would have struck the artist as contravening the logic of composition intrinsic to the Vivian model. The association of the angels with the raising of Eve, seemingly unique to Junius 11, may therefore be derived from the original configuration in Vivian 10v1 in which the creation of Adam is observed by an angel. In the course of adaptation and reworking, the scene came to be reconceptualized as that of Eve’s creation, as argued above. The text at issue—Genesis A 185b hēo wǣron englum gelīce—deserves further commentary. As angels are referred to in the plural (englum dat.pl.) in the text, adhering to the Vivian model provided with a single angel would have been rejected as incongruous. Instead, the artist chose to make an alteration to represent angels in their multiples. While the representation of two angels would appear optimal in view of the corresponding verse referring to Adam and Eve (Genesis A 185b)—as actually and contingently represented in Grandval 5v1 (Pl. 24)—there would have been other considerations to be taken into account in deciding in favor of a host of angels, rather than a pair. Angels are usually represented in groups in Junius 11, although exceptionally appearing in isolation in P46 (Pl.  22)—where a single angel is depicted as an agent of a particular action—or in pairs in Pii and P17—where a pair of angels is represented, together with God, as his closest attendants (seraphim). A simple multiplication of angels, however, would not necessarily lead to two groups facing each other

134 

 5 The Creation of Eve: P9

as represented in P9.91 Rather, a frieze of angels as at San Marco (e.g., Pl. 38) would more likely come into being. Further factors must therefore be specified for the symmetrical grouping actually implemented by more complex geometrical operations. In more specific terms, structural and conceptual in nature, there is the network of correspondences and contrasts at work among a group of pictures in Junius 11. Particularly at issue is P20 (Pl. 12), as discussed in depth from a synchronic–structural perspective in the preceding section. Inasmuch as the primary concern of P9 is evidently with the creation of Eve in paradise, the upper scene in heaven may be regarded as an adjunction of minor importance. In fact, as noted above, the angels figure peripherally as a standard of comparison in the text without implying their presence on the scene of creation. In corresponding fashion, the depiction of Adam and Eve in paradise in the upper half of P20 is of derivative relevance, since the lower scene of hell, the torment and torture of Satan and his company, is focally at stake, as the directly accompanying text testifies (Genesis B 338a–440a; P20 is inserted between verse lines 388 and 389; Table 1, section 1.1). Thus, P9 and P20 are complementary in their treatment of paradise: while paradise is of primary interest in P9, it is of marginal concern in P20. Correspondingly, the heaven scene in P9 and that of paradise in P20, precisely those of peripheral concern in each of the pages, respectively, are supplied to make the two whole pictures fully contrastive with each other down to the organization of their peripheral constituents. In light of the resultant maximal structural opposition, it seems reasonable to assume that the added scenes would have been composed in parallel to the corresponding central ones original to the other pages—the paradise in P9 on the one hand and the hell in P20 on the other. This would entail that the array of angels in heaven would have been introduced to substitute for the single angel in the model (Vivian 10v1) to achieve a fuller contrast with the group of fallen angels in P20, randomly configured. Despite the radical expansion and elaboration executed, however, the original single angel was retained as the central figure in the resulting host, as its unchanged orant gesture eloquently testifies. The hypothesis being advanced here that the hell scene in P20, along with the depiction of the rebel angel army in P3.2a (see below), serves as an impetus for representing the upper picture of P9 by elaborating on the single angel in the model picture also provides an explanatory account of the image of the angel descending the ladder. Since the corresponding model scene in hell in P20 is specifically designed to represent Satan’s agent casting himself through the hell gates in accordance with Genesis B 447a “hwearf him þurh þā helldora,” the artist would have composed, fully concerned with realizing a maximal contrast between P9 and P20, an analogous connection between heaven and paradise and a God’s agent passing it. At this point of conceiving an analogical composition for P9, Jacob’s Ladder may have occurred to the artist by association: “And he

91 Notice, furthermore, that the heaven scene in P9 implies a fully symmetrical configuration, obtained before Eve’s creation, in which God is flanked by five angels on either side.

5.2 A diachronic–comparative perspective 

 135

[= Jacob] dreamed that there was a ladder set up on the earth, the top of it reaching to heaven; and the angels of God were ascending and descending on it” (Gen. 28:12; Coogan 2010; cf. Raw 1976: 140–141; Karkov 2001: 57–58; Hilmo 2004: 90). More specifically, Adam sleeping at lower right may have evoked Jacob sleeping, whose vision would in turn have made its way into the upper zone through analogical adaptation (cf. Raw 1976: 140).92 It should be emphasized, however, that Jacob’s Ladder on its own hardly would have played a decisive role in the creation of the upper scene in P9, notwithstanding the contentions often voiced. It is primarily the intervisual opposition to P20 that critically would have been in force behind it. As Raw (1976: 140) asserts, the heaven scene in P9 is unparalleled elsewhere and unsupported textually in Genesis A and Genesis B (see also Broderick 1978: 168). In the complete absence of parallels to reliably draw on, it should best be inferred that the scene at issue was of the Junius artist’s (or his predecessor’s) own making, urged on intervisual grounds that were internal to his own work. The uniqueness of the heavenly scene of the host of angels as part of the creation of Eve thus ultimately should be ascribed to the distinct treatment of the Fall of the Rebel Angels in the Old English and Old Saxon reworking of Genesis. Without such relevant materials to be contrasted with, it is unlikely that there would have been inspiration and enticement encouraging the artist to elaborate the model’s figure of the prayerful angel in the manner he actually chose. One may then generalize the intervisual dependence involved in simple terms, as follows: where there are no hell gates depicted, there are no heaven gates likely to be represented. Analogously, the temptation of Adam by Satan’s agent, the scene that took place immediately after the latter’s arrival in paradise, was added as a complement to the hell scene—although this juxtaposition must visualize the scenes far earlier than their textual references (454a–459b for the right picture of P20, and 491a–494b for the left one; for details, see section 7.1 below)—to make a perfectly balanced landscape of paradise with P9, in which God created Eve out of a rib taken from Adam. Each of the two pages is thus augmented with a secondary component picture whose original counterpart is of focal presence on the other page. Also arguably responsible for the army of angels in P9 may have been P3.2a, in which Lucifer is attended on both sides by his adherents holding palms as symbols of victory (section 3.1). In contrast to P20, Lucifer and his company here are shown standing at the zenith of their power in heaven. The major contrasts involve Michael, God’s current deputy versus Lucifer, God’s previous deputy, and God’s followers in heaven versus God’s enemies in heaven. Accordingly, on a less conspicuous scale, P3.2a constitutes an opposition to the heaven scene in P9, as both armies are commonly located in heaven and arrayed in a comparable manner. By virtue of the closer conceptual and

92 For the extra-biblical, apocryphal background to the sleeping Adam and his vision of heaven, see Broderick (2009: 393).

136 

 5 The Creation of Eve: P9

formal resemblances thus in evidence, P3.2a may have served as a more accessible visual model for the upper section of P9. Finally, one may wonder why the artist took the trouble to reuse the image of the animation of Adam as part of the scene of the creation of Eve, rather than use it straightforward for the creation of Adam. This question actually falls into two parts: first, the use of the image in question inseparably in conjunction with the other two scenes, and second, a separate use (or disuse) of this image for the creation of Adam. As it is, the first part of the question has already been answered by identifying the scene of God taking the rib from Adam as an initial motivation and the textual reference to angels as a further inducing factor, invoking then the integrity of the three images under discussion as a compositional framework, and finally arriving, as a deductive consequence, at the reworking of the animation of Adam as that of Eve. The second part of the question concerns the artist’s treatment of the creation of Adam, more specifically, his appropriation of the leftmost image of the animation of Adam in Vivian 10v1 to represent the animation of Eve. How did the artist visualize the creation of Adam in Junius 11 at all? Did he use the same model twice for two different subject matters, or did he turn to another model for the sake of diversification? This is a highly speculative query, given that the picture is absent in the current manuscript. However, it is well worth asking and requires a serious answer, if by necessity conjectural in content. The corresponding illustration doubtless must have been included in the three lost leaves at the beginning of the second quire or between the extant pages 8 and 9 (Doane 2013: 5–10; Table 2, section 1.1). It is quite unlikely that an episode of such paramount significance simply escaped visualization, nor is it plausible that the missing picture is limited in its coverage to the animation of Adam. On the contrary, given the illustration for the creation of Eve, it seems highly probable that the story would have been illustrated at least as lavishly and exquisitely as its Eve counterpart. Obviously, however, Vivian 10v is hardly rich enough as source material for this purpose, as only a single relevant image is available there. As realized in a fuller form in San Paolo (Pl. 34), San Marco (Pls. 38, 40), and Millstatt (fol. 3v, Pl. 58; fol. 6r, Pl. 59), the creation of Adam would have comprised two episodes in the Cotton Genesis tradition (Kessler 1977: 13–17)—the formation and the animation of Adam. Accordingly, it seems unlikely that the leftmost image of the animation of Adam, as is realized in the Vivian page, would have been adopted as it is, all by itself. Rather, the artist would have sought a more elaborate model of visualization than is available in Vivian 10v. A most probable model instead would have been offered by a Touronian predecessor of the San Paolo Bible, the hypothetical manuscript from Tours that is featured as ancillary but significant sources for Junius 11 illustrations (sections 2.2, 3.2, and 4.2). It would, however, be futile to add further speculation on this issue impervious to empirical verification, other than conjecturing on a final note that, inspired by the Vivian and Grandval Bibles among others, a group of angels would have been featured no less prominently in the representation of Adam’s creation (see also San Marco and Millstatt; Pls. 38, 58, and 59) than in the illustration for the creation of Eve (P9), and that some formal operations may have been

5.2 A diachronic–comparative perspective 

 137

implemented on the pair of Adam and God to achieve a more effective reconfiguration. In any event, the top leftmost pair of God and Adam in Vivian 10v1 still would have been left unexploited and thus stood ready for adaptation when the artist turned to engage in the creation of Eve after finishing that of Adam. While the postulation of the tripartite representation in Vivian 10v1 consisting of the Animation of Adam, the Observing Angel, and the Removing of a Rib from Adam as an integral model for P9 has been substantiated by the principled explanations that it offers for the motivations and mechanism of adaptation including all major derivational specifics involved, there are still some residual issues awaiting credible accounts. Most recalcitrant among the unsolved problems are those concerning God’s figuration. While God holds nothing in Junius, he carries a staff in his left hand in Vivian. Given that God almost always holds a scroll/scepter in Junius, its total absence in P9 looks extraordinary, particularly when the alleged model—Vivian 10v1—depicts a corresponding item in God’s hand. Partly related to this point, in the scene of the Drawing a Rib, God touches Adam only with his right hand in Vivian, whereas he uses both hands in Junius. Moreover, God’s bending posture differs slightly but appreciably: he bends more closely and tenderly over Adam in Junius. In the Formation of Eve scene, God contacts Eve by the hand in Junius, but he keeps distance from Adam in Vivian in the original scene of the Animation of Adam. Finally, God stands in a higher position than Eve in P9, whereas he and Adam are positioned at the same ground level in Vivian 10v1. These fine differences defy straightforward explanations insofar as one adheres to the strongest version of the claim that P9 is based on Vivian 10v1, namely, that P9 would have been modeled exclusively on Vivian 10v1, with no other sources involved. Since such a maximally strong interpretation proves not to be fully feasible in light of a handful of issues unaccounted for, one needs to weaken the original claim to the extent empirically due by conceding that P9 would have been shaped complementarily by a concurrent model, which may be held responsible for the intractable problems raised above. In fact, there is such a secondary model available, but with a twist. At issue is a mirror image of Bamberg 7v2, more precisely, a reversal of the two left pictures in the second register, the one depicting the Drawing a Rib from Adam and the other for the Introduction of Eve to Adam (see Fig. 3—adapted—in comparison with Pl. 23—original). In the mirror representation, the Drawing and the Introduction scene come to stand at the right and the left, respectively, with the two figures of God in the inner sides, separated from each other by a large tree in the center. The emergent overall configuration perfectly matches the composition of P9, with the tree substituting for the ladder. With respect to the details of God’s figures, he holds no items in his hands in Bamberg in exact parallel to Junius. Thus, the peculiarity of God holding nothing in the particular picture in Junius may be credibly attributed to a carryover phenomenon from the Bamberg source. Moreover, God’s hand gestures in both scenes in Bamberg are identical to those in Junius: God touches Adam with both hands (Fig. 3, right), and makes direct contact with Eve (Fig. 3, left). Stooping deeply, God’s bending posture in Bamberg corresponds better with that in Junius. God stands at a higher level than Adam and Eve in the Intro-

138 

 5 The Creation of Eve: P9

Figure 3. Mirror image of Bamberg 7v2 (left half). Bamberg, Staatsbibliothek, Msc. Bibl. 1 (Bamberg Bible), fol. 7v (detail, adapted), Genesis frontispiece (licensed under CC BY-SA 4.0).

duction scene in Bamberg. By invoking the mirror image of Bamberg 7v2 (Fig. 3) as a source, then, one may offer principled accounts for these specific correspondences that otherwise remain beyond understanding.93 Of further interest in this connection is the Creation of Eve at San Marco (Pl. 43). In it two scenes are depicted side by side, the Drawing of a Rib from Adam at the left and the Formation of Eve at the right, with a left-to-right narrative progression. With the two figures of God appearing back to back and close to the center, this constitutes an exact mirror image of P9. The direct involvement of the Formation of Eve, in distinction from the following scene of the Introduction of Eve to Adam as in Bamberg 7v2, makes the San Marco image the closest match with P9 in conceptual terms without the need to execute a scenic conversion. With the overall conceptual identity involved, then, the San Marco picture might be regarded as an ideal model that would obviate the role that has been attributed above to Bamberg 7v2, an analysis that necessitates a conceptual conversion in the course of adaptation—although slight—from the Introduction to the Formation of Eve. On closer inspection, however, this allegedly maximal conceptual correspondence between P9 and San Marco is compromised by disagreements with P9 in specific details that disappear under the postulation of Bamberg as an alternative auxiliary model. Concretely, the figure of God in the Drawing of a Rib from Adam at San Marco holds a

93 Of related interest, the sloped ground in P9, as opposed to the flat surface in Vivian 10v, may be viewed as a further reflection of Bamberg 7v.

5.2 A diachronic–comparative perspective 

 139

cross-staff, as in Vivian. Accordingly, he touches Adam only with his right hand, in parallel to Vivian. Moreover, God bends only slightly over Adam, much as in Vivian. God occupies the same ground level as Eve in the making. Therefore, insofar as one sticks to San Marco as a secondary model, to the sacrifice of Bamberg, these discrepancies remain unsolved, which would, in turn, cast doubt on the legitimacy of positing San Marco as an explanatory model in place of Bamberg. The incapability of accounting for the crucial issues that would otherwise be subject to explanation thus may disqualify San Marco as a formative source for P9. At this point, it may be necessary to explain why the artist took the trouble of using the mirror image—a virtual representation conceived in the artist’s mind, rather than the actual one externalized in his presence—of Bamberg 7v2. This fundamental issue will become explainable by setting it in the interpretive framework predicated on the Vivian model. As may be recalled, the triad of pictures at the upper left corner of Vivian 10v1 requires arranging the two successive events of the Drawing a Rib and the Forming Eve in a right-to-left sequence. It is this foundational mapping that would have given impetus to conceptualizing the corresponding scenes in Bamberg by reversal, the earlier phase placed to the right of the later one. Such a formal operation would have harmonized optimally the two variant representations of the same episode—Vivian and Bamberg—and hence maximized their value as viable models for P9. However much explanatory power the Bamberg model may be asserted to be equipped with, it is only secondary to the Vivian exemplar, as should be partially clear from the preceding paragraph. The Bamberg model—its mirror version, to be more exact—makes sense only in the framework that has been determined by the Vivian counterpart in advance. Bamberg is thus subordinate to Vivian as far as the source materials for P9 are concerned. Bamberg’s dependence on Vivian, however, comes to light most strikingly with regard to the watching angel, whose presence counts as one of the defining features of P9. Whatever model one may conjure up as a prime basis for P9 must contain an explicit representation of an angel. In the absence of an angel as an inherent figure in the Bamberg model, it is unqualified as a primary inspiration for P9. It presupposes some other model like the Vivian imagery for its own viability. In conclusion, Bamberg 7v2 would have served as a complementary, secondary model, while Vivian 10v1 set the stage as a primary basis for P9. Only by postulating the close coordination of these two Touronian materials can one explain all significant features of P9 in a principled way. Finally, one may pose another question of the fundamental order, still to be answered: Why did the artist bother to use a complementary model, Bamberg 7v2 (left half, horizontally reversed), in addition to the Vivian material? It would have been far simpler to stick to a single model than to make a synthesis of two different sources. One advantage, if nothing else, of the Bamberg model over the Vivian counterpart is its conceptually closer approximation to P9 as a depiction of the Creation of Eve. On the left side scene, Eve is commonly featured in P9 and Bamberg, whereas Adam figures instead in Vivian. This invariant identification of Eve obliterates the identity change

140 

 5 The Creation of Eve: P9

that would otherwise be necessitated, which definitely would have made a great asset of the Bamberg image, together with a symmetrical arrangement there wherein God stands closest to the center on both halves of the panel. Thus, despite the complexity incurred in execution, the concurrent use of the Bamberg model has inherent merits that would have convinced the artist to adopt it eventually in complementation to the primary Vivian source.94

94 Compositionally similar to P9 to some extent is the first panel on the Hildesheim door (Kaspersen 1988: 82, fig. 4). Apparently ambivalent with respect to the identities of the two human figures depicted, the picture is open to competing readings (see below). In purely formal terms, however, the panel might be regarded as derivative of—or at the very least formally parallel to—P9, through the intermediary of a few transformations operated on P9 in chain reaction, apart from further minor adjustments in detail: Relocate God/Adam from the far right to the center, reposition Eve/God to the right of God/Adam as made available, remove the repositioned figure of God at the far right (presumably for lack of space for representation), and shift the angel to the far left, now vacated by the repositioned Eve/God. Underlying the sequence of these movement transformations is a reconstitution of left-to-right narrative progression. As noted above, the first Hildesheim panel poses recalcitrant problems in interpretation (Broderick 1978: 158–162; Tronzo 1983: 357–358; Kaspersen 1988: 81–82): the two human figures involved are each susceptible to identification as either Adam or Eve. Logically speaking, then, there are four different combinations available: (i) left = Adam, right = Adam; (ii) left = Adam, right = Eve; (iii) left = Eve, right = Adam; (iv) left = Eve, right = Eve. Interpreting both humans as Adam, the first option is dominantly held (e.g., Kessler 1977: 16–17; for bibliographical details, see further Broderick 1978: 158n40; Tronzo 1983: 357n3; Kaspersen 1988: 98n17). In this view, the image as a whole constitutes a conflation of two scenes of the Creation of Adam, the formation and the animation of Adam—as fully depicted in San Paolo 8v1 (left)—with God’s second role undepicted at Hildesheim. Accordingly, the right tree divides the picture into two successive scenes progressing from left to right, again as in San Paolo. To rephrase in more revealing terms, the left scene concerns the Creation of Adam in progress, whereas the right focuses on the final result—the animated Adam—with a concomitant suppression of the agent—the Creator—from representation. Option (ii) is embodied by Tronzo’s (1983: 362–366) interpretation, according to which the Hildesheim image closely corresponds to P9 iconographically as well as compositionally. In fact, Tronzo attributes such close correspondence to Anglo-Saxon influence. The third option is proposed by Kaspersen (1988: 92–96): the left scene represents the Creation of Eve, which is at the same time visioned by Adam. Reading (iv) may be dismissed merely as a logical possibility with minimal plausibility, unlikely to be encountered in the literature.

6 Adam and Eve in Paradise: P10/P11/P13 6.1 A synchronic–structural perspective P10 (Pl. 7) depicts the scene in which God introduces Adam and Eve to paradise, blessing and admonishing them. It hardly would be necessary to make a categorical distinction between the two speech acts of blessing and admonishing; most importantly, the couple is introduced to life in paradise, on which occasion God addresses the couple in welcoming words. In this connection, notice that in Junius there is no scene of God introducing Eve to Adam, as observed above. Thus, P10 serves as a substitute for such a scene, as can be found in Bamberg, Grandval, Vivian, and San Paolo. Then, having completed all his work of creation, God beholds all his creations from above (P11, Pl. 8). As the final picture in the three-part series, P13 (Pl. 9) illustrates life in paradise, with a focus on the couple, after God has returned to heaven and let all creatures make their living on earth. While depicting similar scenes of Adam and Eve in paradise as outlined above, P10 and P11 are polarized in compositional terms along a number of parameters. The remaining member, P13, stands broadly at a midpoint between these two extremes. Moreover, the contrast between P10 and P11 is parallel to that between P2 and Pii (section 2.1). P10, like P2, is horizontally aligned, whereas P11, like Pii, is vertically oriented. In other words, the figures are concatenated along a single horizontal axis in P10, forming a single tier, whereas they are placed on top of another in P11, thereby forming a triple-tiered structure—God, Adam and Eve, and animals from top to bottom. In regard specifically to the individual figures, God turns left in P10 but faces the front in P11. Adam and Eve face in the same direction toward God in P10, but are arranged in confronting position, both looking up at God in P11. P10 is relatively simple in architectural framing, while P11 is more elaborate, as a comparison of particularly the roofs, towers, and windows makes evident. P10, like P2, is an interactive scene (interactive with his creations, i.e., humans in P10 or angels in P2), while P11, like Pii, is static (beholding and contemplating from above, without acting on others). The interactive mode or lack thereof naturally correlates to the distance between God and the couple—proximate in P10 as opposed to distant in P11. While God is standing in P11 as well as P10, the lack of dynamicity or a lesser degree of dynamicity in P11 as opposed to P10 may be ascertained in the use of the footstool (scabellum) on which God stands in P11 but not in P10. Both P11 and Pii are provided with a larger space, a full page or a close approximation to it, in comparison to P10 and P2, respectively, which are only half-page illustrations. With the two polar opposites thus differentiated, the remaining member of the group, P13, stands at their intersection. The symmetrical arrangement of Adam and Eve, the multiple tiers for accommodating figures—God, the human couple, animals, and sea creatures—and the lack of interaction between God and humans may group this page with P11. The profile view of God and the half-page allocated for illustration, however, make it correspond to P10. The failure of P13 to fully align with P11 seems hardly surhttps://doi.org/10.1515/9783110788068-006

142 

 6 Adam and Eve in Paradise: P10/P11/P13

prising in light of the major concern of this page. As noted above, P13 may be derived from P11 through its focus on creatures in paradise in general and Adam and Eve in particular, and a concomitant defocusing on God, who is back in heaven. Put another way, while P11 is concerned equally with God and his creations on earth, P13 concentrates on the latter, thereby relegating the former to the background. In accordance with such a thematic shift, there is no longer compelling motivation to represent the prototypical frontal figure of God in a maximal format. P13’s focus on paradise almost to the exclusion of the heavenly sphere results in a large upward expansion of the pictorial zone of earth with the corresponding reduction of God’s celestial domain. Concomitant with such a radical reallocation of space is the replacement of the hexagonal enclosure with double arches. The appearance of such a complex structure rather than a simple semicircular arch, as in P10, may be subject to explanation. This geometry may be conceptualized as a reflex of the lower half of the hexagon in its outline whereby the bottom horizontal line in the center is linked on both sides to upward diagonal lines. This geometrical shape would appear to be largely analogous to the spandrel, the central part of the double-arch frame where the two arches are joined. Furthermore, the few window-like decorations on the spandrels in P13 may be viewed as vestiges of the windows of the buildings adjacent to the enclosure in P11. In short, the original hexagon would have been reduced by half, whereby the upper half was deleted and the lower counterpart took its place, incorporated a few traces of the buildings, and was curved in its finish as appropriate.95 The mandorla framing God in profile in P13 is peculiar in its precise shaping: interior to the outer frame in the right half can be found another, thicker arc, the overall figure being reminiscent of a bowl. Thus, it looks as if God is contained in a vessel. With rays of light emanating out, the whole configuration may evoke a vessel of light held by an angel in the creation pictures, P6 and P7.2 (section 4.1; cf. Broderick 1978: 113). A clockwise rotation by 90 degrees, particularly of the vessel in P6, which is depicted in greater detail than in P7.2, may bring to light a formal analogy between the vessels in P6 and P13. Interpreted this way, the figure of God encased (P13) may be characterized as transitional between the configuration of God and the bowl overhead (P6/P7.2), and the representation of God enclosed in a mandorla (P7.1). These related configurations are embodied by the varying similarities to the physical shape of a bowl along a scale from concrete to abstract. In this light, Raw’s (1955: 318) characterization of the bowl of

95 That P11 and P13 are variations of each other has long been known, as the following quotations confirm: “The drawing on page 13 . . . is closely related to that on page 11, in fact, the artist probably derived the basic figure unit, that of Adam and Eve flanking a centrally placed tree, from the same source as the drawing on page 11” (Broderick 1978: 192); “. . . as Barbara Raw has noted, that the composition on page 13 is almost identical, at least in the figure types and poses, to that on page 11, . . .” (Broderick 1978: 189). Emphasizing the similarity so excessively, however, has failed to address squarely the exact ways in which their resemblance and conversely their distinctions are undeniably in evidence.

6.1 A synchronic–structural perspective 

 143

light as a variation of the mandorla has a broader basis for empirical support than she originally conceived. The insight presented above that P11 and Pii are organized in close parallel on the basis of the analogous structural variation with their respective opposites P10 and P2 will lead to a better understanding of their figural and compositional details in mutually illuminating ways. The hexagonal enclosure surrounding God in P11 seems analogous in contour to the seraphim standing just behind God in Pii. As argued at length in section 2.1 above, God and the seraphim are fully integrated into a unified whole in Pii; thus, the surrounding seraphim may be viewed as a contour of God himself. In this light, the architectural enclosure in which God is implanted in P11, notably its hexagonal contour line, may readily lend itself to an explanatory account. At the same time, the upper half of the enclosure finds a close analog in the angled arch, arguably another echoing of the seraphim in Pii. The cherubim under the feet of the throne in Pii, corresponding to the base of the hexagon in P11, may be compared with Adam and Eve just below the hexagonal frame in P11. Although the orientation of the faces differs, their location relative to God’s feet is strikingly similar. Adam and Eve’s two hands raised to shoulder-height in P11 are very much reminiscent of the widespread wings of the cherubim in Pii. The overlapping part of the flanking cherubim directly underneath God’s feet corresponds to the central tree intervening the couple. While God is seated in Pii but standing in P11, the manifest presence of the footstool on which he stands in the latter picture specifically links to the figure seated on the throne, since the standing figure of God is seldom accompanied by such a device. The unusual occurrence of the footrest under the standing figure of God in P11 rather than somewhere else therefore may be adduced as eloquent witness to the intimate relationship between Pii and P11 as claimed here. Finally, the pictures’ complementarity in their domains of interest may account for the polar contrast in focal zones of illustration. The focus on God’s absolute presence—God and his immediate surroundings—in Pii naturally leaves the lower part of the panel unarticulated in a cloud-like state. By contrast, the central concern with God’s act of beholding his creations on earth gives exactly this part a pronounced presence at the expense of his immediate exteriors in heaven. While similar figural and compositional correspondences are expected to obtain between the simpler variants P10 and P2, they are likely to be far less spectacular on account of the smaller number of features available for exploitation that are constitutive of the relatively plain pictures involved. Of vital interest here is the large apparently blank space between God and Adam in P10. In other words, the two figures are placed apart, concomitant with a central space of the page left vacant. A comparison with P2, however, provides insight into this seemingly weird disuse of the central zone. Exactly in the corresponding space in P2 is found a seraph standing to the left of God. Since a seraph is not featured at all in P10, the corresponding spot is left unfilled, with Adam and Eve substituting for the two angels including Lucifer in P2. Still more significant in this connection is the finding that another human figure, dry-point sketched, emerges

144 

 6 Adam and Eve in Paradise: P10/P11/P13

under ultraviolet light in the blank space at issue (Broderick 1978: 174–177), although one can see it to some extent, nonetheless, in the high-resolution image available online. This figure stands and assumes the same pose as the actual figure of Adam to the left. This discovery shows that the artist found fault with his initial attempt at picturing Adam in the center. In thematic terms, however, there would be nothing conceptually wrong about this original arrangement. On the contrary, one may go so far as to assert, it would otherwise have been optimal, all things being equal. As it turns out, however, constrained by the specific artistic framework he was committed to, the artist made a fully accountable decision that deviated from the purely logical ceteris paribus argument. Conceptualizing P10 as a structural analog of P2, the artist believed it in his best interest to preserve the model’s essence in its formal reproduction in figural and compositional terms. Thus, he cancelled his earlier design and shifted the figure of Adam to the left to secure the critical space between God and Adam out of respect for the corresponding arrangement in the model picture, P2. The disuse of the early sketch, leaving the space blank in the final version thus convincingly demonstrates how much value the artist placed on the resultant space secured between God and Adam, despite all logical thinking pointing otherwise. This seemingly irrational decision, however, was fully motivated: this is exactly the space that is occupied by a seraph in P2, an alien element now in P10. Accordingly, the relative positioning of God and Adam/Eve must conform to that of God and the two angels, to the exclusion of the seraph. Therefore, the rejection of the original plan in favor of the modified version constitutes eloquent testimony for the claim that the artist had a keen awareness of the compositional parallelism between P10 and P2 and tried his best to realize it in execution. On the face of it, P10 differs from its presumed analog, P2, in its accommodation of an architectural frame. This apparent disparity, however, by no means contradicts the claim that P10 is largely inspired by and modeled after P2. The superficial lack of an architectural frame in P2 would have been no hindrance to providing one in P10. Since P2 is a simplified variant of Pii, and since Pii as an elaborate version of P2 is equipped with an architectural frame, its presence is implied, though not manifested, in P2. Thus, the latent presence of framing in P2 may guarantee compatibility of P10 with an explicit representation of such a structure, although naturally enough the absence of a manifest form would have been as acceptable. As detailed above, P11 depicts the scene of God beholding his creations from above. This naturally evokes P7.1 by association, which represents God observing his creations on land rather than God being engaged in creation (section 4.2). It may be of interest then to compare the two conceptually related pictures with a view toward obtaining mutually supporting and enriching interpretations of them. On the whole, P11 may be viewed as a continuation of P7, wherein God beholds his creations as in P7, but Adam and Eve are featured this time, immediately subsequent to P10—the facing page—where God welcomes Adam and Eve to paradise. The thematic continuity finds its formal counterpart in the same vertical orientation as with P7.1. The hexagonal architectural frame enclosing God in P11 may be characterized as a variant of a mandorla in P7.1, as pointed

6.1 A synchronic–structural perspective 

 145

above. It may be of interest to note the trees grown large with the lapse of time intervening between P7 and P11. Notwithstanding the thematic continuity running through them, the two pictures may be characterized as largely contrastive in their manners of representation and organization—elaborate/naturalistic (P11) versus simple/schematic (P7.1)—as embodied by the following properties. While P7.1 is represented sparsely with minimal figures, P11 offers a thick, figure-rich depiction. This contrast is expressed most vividly by the distinct treatment of animals and birds. P7.1 offers a single instance of each, whereas one finds P11 crowded with multiple examples. Moreover, a single row of animals and plants in P7.1 looks highly schematic and abstract, whereas the mixed crowd of these creatures in larger numbers in P11 fits the natural setting. The modest depiction characteristic of P7.1 is further demonstrated by a large blank space in the upper portion of the lower dome—the firmament—below the standing figure of God. By contrast, there is no such blank space left underneath the hexagonal enclosure in P11. The contrast between elaboration and simplicity is manifested in the form of the frame as well: an abstract simple arc (P7.1) on the one hand and a crenellated hexagonal architectural construct (P11) on the other. An elaborate illustration requires relatively larger space (an almost full-page drawing for P11), whereas a smaller space (a half-page illustration for P7.1) is sufficient for a moderate representation. Another outstanding difference between the two otherwise comparable pictures is one of symmetrical and asymmetrical arrangement. P11 manifests symmetry in a clustering of figural and configurational features, including the figure of God in frontal view, the couple standing face to face, the regular configuration of the trees, and the hexagonal enclosure. In contrast, P7.1 offers asymmetrical arrangements instead, such as the figure of God in three-quarter view and the absence of conspicuous figures positioned symmetrically like Adam and Eve in P11. While the semicircular frames are symmetrical by nature, their abstract, purely geometrical existence may scarcely render the whole picture as symmetrical in appearance as is P11. This contrast in symmetricality is concerned not so much with representation of individual figures as with overall pictorial organization. The predilection for symmetry in P11 and the lack thereof in P7.1 may accordingly be subject to understanding by relating them to the respective groups of pictures of which they are members. As treated in depth in section 4.1 above, P7.1 is one of a three-part series of pictures with P6 and P7.2. Comparably, P11, together with P10 and P13, is subsumed under one group. At first glance, then, P11 and P7.1 might seem fully parallel in the ways in which they are structured at a higher level of organization. Upon closer inspection, however, they prove to be far from analogous. The three creation pictures are organized exactly the same way. Where they differ are in the contents themselves, specifically in the particular creations God accomplished on each day. Most importantly, the three distinct scenes involved are pictured invariably from the identical perspective. By contrast, the members of the other group differ not so much in their contents—they all represent God, Adam, and Eve as major figures in one way or another—as in their perspectives

146 

 6 Adam and Eve in Paradise: P10/P11/P13

and points of emphasis. These varying conceptualizations are linked to the differing ways of composition and configuration in visualization. Specifically at stake is the variation between symmetrical and asymmetrical modes of representation, whereby P11 is defined as symmetrical in distinction to P10, with P13 standing at their intersection, as detailed above. Thus, P11, to the exclusion of P7.1, is implicated in the variation in terms of symmetry/asymmetry. In accordance with the structural underpinnings predicated on interpictorial organization, then, P11 is distinguished from P7.1 by its strong attachment to symmetry, despite their broadly shared concerns with the same subject. While depicting God’s creations on earth, P11 and also P13 are concerned exclusively with living creatures—animals, birds, fishes, and plants—as opposed to the sun, the moon, and the stars. In this light, the blank dome in P7.1 would be perfectly intelligible for the same reason P11 and P13 are without an explicit representation of celestial bodies—also of God’s making—on the assumption that P11 and P7.1 are variations of each other sharing a common purpose of visualization. Indeed, these two variant pictures are differentiated by the presence (P7.1) versus the absence (P11) of a blank space between God and his creations on earth. Yet this empty expanse in P7.1, as opposed to a total lack of the celestial sphere as in P.11, can be readily understood. Since the Junius artist was primarily committed to maximizing an overall compositional coherence with the immediately preceding two scenes of P6 and P7.2 due to their thematic continuity and integrity, he had no choice but to fully reproduce the sky—the upper zone of the lower dome—and leave it blank, as argued in section 4.1. Noteworthy further is the depiction of fishes in P13 in contrast to their absence in P11, a difference that might simply have been a matter of available space. In contrast, the representation of the water but the failure to depict fishes in it in P7.1 possibly might have been simple negligence, but more likely—in light of the deliberate omission of the heavenly bodies—may have been the result of rational design and decision-making by letting the depicted water imply the existence of water creatures in it, out of respect for the conventional limitation to a single row of creatures for representation (section 4.2). On another front, P11, God beholding his creations on earth, is in opposition to P17, God beholding his enemies in hell. Outstanding above all, however, is the parallelism in overall design and composition between the two illustrations. Framed by the fairly detailed architectural structure with towers on the sides, the upper half of the page is occupied by the frontal figure of God at the center, gestured almost identically, except for the difference between standing (P11) and sitting (P17). Yet the footstool specifically depicted in P11 approximates the standing figure of God to the enthroned counterpart, as observed above, and may make the distinction in question more apparent than real. Although complementarily distributed (above vs. below in the panel), the crenellated enclosures constitute another conspicuous similarity. Interestingly enough, however, the two crenellated enclosures, enclosing God like a mandorla on the one hand (P11) and containing the devils in hell on the other (P17), are structured differently, despite their superficial similarity at first sight. The enclosure in which God is situated is open

6.2 A diachronic–comparative perspective 

 147

to paradise, as suggested by the uncrenellated line dividing God in heaven and Adam on earth. In contrast, the hell counterpart is completely sealed off from heaven. Of particular interest is the sawtooth crenellation interior to the enclosure bordering heaven. The absence of the seraphim in P11 may be explained as a substitution of the enclosing structure for the attendants, as argued above in comparison with Pii. The overall analogous shape of the two otherwise distinct entities—compare P11 (enclosure) with Pii (seraphim), the elaborate version of P17 (section 2.1)—may be adduced as supporting evidence for this claim.

6.2 A diachronic–comparative perspective By far the closest to P10 in compositional terms is Grandval 5v2 (right, Pl. 24). To a lesser extent, it also corresponds to the Hortus Deliciarum, fol. 17r (lower right, Pl. 68; Green et al. 1979: pl. 10, no. 18 and fig. 21). Despite general resemblances, P10 stands out from these two images in notable respects. First, while God appears to the left of the couple both in Grandval and the Hortus Deliciarum, the reverse arrangement obtains in P10, with God standing at the right. This alteration in relative positioning seems hardly surprising, given that Junius exhibits marked preferences for placing God on the right side, as pointed out repeatedly above. Second, unlike these presumable cognates—Grandval 5v2 and the Hortus Deliciarum 17r—Adam and Eve are raising both hands in P10. In Grandval 5v2 and the Hortus Deliciarum 17r, Eve is lifting only one hand, while exclusively in the Hortus Deliciarum the Creator takes one of Adam’s hands. The raising of both hands to shoulder-height recurs throughout the group of P10/P11/P13, thereby pointing to the integrity of the three pictures in question. Third, the enclosing in an architectural frame is a unique feature of P10, unknown to its Grandval and Hortus Deliciarum counterparts. This may be attributed to the variation with P11, which, because of the parallelism to Pii as substantiated in the preceding section, is inherently associated with such a structure. Thus, in the absence of an architectural frame in the two cognates of P10 in Grandval and the Hortus Deliciarum, the unique occurrence of one in P10 should best be derived from its elaborate variant, P11, which, as verified independently by its structural analog (Pii), contains such a framing structure as one of its inherent properties. Of the major pictorial components that constitute P11, the motif of God blessing in frontal view and that of Adam and Eve facing each other with a central tree in between are seen pervasively. Because of such a ubiquitous presence and generality in their symbolization, they can hardly provide a useful guide for identifying concrete sources for P11. By contrast, the crowd of animals seems specific enough both in content and distribution to serve the present purpose. Accordingly, using the depiction of animals in a large group as a heuristic criterion for specifying sources, one may arrive at the suspicion that the following four pictures may possibly have provided inspiration for P11: Bamberg 7v1 (right, Pl. 23); San Marco (Pl. 42); Millstatt 9r (Pl. 61); and Utrecht 82v (Pl. 81).

148 

 6 Adam and Eve in Paradise: P10/P11/P13

Affiliated with the Cotton Genesis family, the first three works all represent the same scene, the Naming of the Animals (Kessler 1977: 23). This scene is far from extensively illustrated in the Cotton/Touronian tradition. Utrecht 82v is a visualization of Psalm 148, a call for praising God from all quarters in heaven and on earth. Comparisons of these four possibly related pictures with P11 and P13 may be made with regard to several features that, taking on varying values across the group under examination, are of significance in characterizing each member and determining its probable contributions to the identities of the two Junius 11 images. The following five features will be explored below: (i) left versus right orientation; (ii) frieze-based (linear or one-dimensional) versus plane-based (two-dimensional—implicated in both horizontal and vertical extension) arrangement; (iii) symmetry versus asymmetry; (iv) singles versus pairs; (v) treatment of birds. These features crisscross the divisions based on the pictorial contents (e.g., the Naming of the Animals) and/or on the genealogical affinity (e.g., membership in the Cotton Genesis family/the Touronian Bibles). The animals may be divided into two classes according to their horizontal orientation, leftward versus rightward. The three Cotton members—Bamberg, Millstatt, and San Marco—all exhibit a marked preference for leftward orientation, as the following counting of relevant animals indicates: in Bamberg, fourteen instances of leftward-facing creatures versus three instances of rightward-facing creatures; in Millstatt, six leftward versus zero rightward; in San Marco, twelve leftward versus four rightward; in P11, five leftward versus five rightward; in P13, five leftward versus five rightward; and in Utrecht, seven leftward versus twelve rightward. By contrast, Utrecht evinces a seemingly contrary pattern of favoring the rightward orientation. The difference between the two opposite distribution patterns—Bamberg/Millstatt/San Marco versus Utrecht—proves to be statistically significant (p < 0.001, by Fisher test). While the strong predilection for the left undoubtedly may be counted as a significant property common to the three Cotton members, it is not legitimate to define Utrecht by the opposite characteristic, as shown immediately below. All that may be said in rigorous terms is that Utrecht 82v does not share the same preference for the left orientation. Since Adam is featured as the name-giver in the three depictions of the Naming scene, it is appropriate that the animals to be named should be aligned facing toward Adam to receive names. This expectation holds true, as has been just substantiated: in all three pictures, with Adam standing at the left, the leftward-facing animals prevail over those oriented in the opposite direction. By contrast, given that Utrecht 82v has no bearing on the Naming of the Animals, it follows as a matter of course that no particular preference for horizontal orientation is in effect in Utrecht 82v, other things being equal. Much the same reasoning may apply fully to P11 and P13 as well. Since these Junius illustrations are concerned with depicting animals living in paradise, there should be no reason why they favor one directionality over the other, again other things being equal. Accordingly, their horizontal orientation must be of no significance at all. More specifically, while Adam, along with Eve, appears in the picture, it is not in the capacity of a name-giver here. In fact, the couple has no particular relation to the animals in

6.2 A diachronic–comparative perspective 

 149

this context. Comparing then the left/right distribution of 10:10 (for P11 and P13 as given above) with 32:7 (for the three Cotton members), one gets a contingency table [10, 10, 32, 7]. Running a Fisher test provides a p-value of 0.015, meaning that the distribution pattern involved is statistically significant. Accordingly, Bamberg 7v1, Millstatt 9r, and San Marco exhibit biases toward left orientation, as observed above, in contrast to the sum of P11 and P13. In the Junius pictures, the left orientation is as likely to obtain as the inverse. At this point, a comparison between P11/P13 and Utrecht 82v may be in order. Performing a Fisher test on a contingency table [10, 10, 7, 12] gives a p-value of 0.523. Therefore, one may conclude that there is no statistical difference between the two distributions. More concretely, much as P11/P13 is neutral in horizontal orientation, Utrecht 82v is unbiased, lacking significant preference for either direction. In all five pictures except P11, the entire herd of creatures represented consists of several rows of friezing animals and birds: three layers in Millstatt, San Marco, Utrecht, and P13—including the water creatures—on the one hand, and as many as five in Bamberg on the other. Moreover, these row-based groupings are arranged overall in largely pyramidal form, whereby the lower a level is, the more animals it contains. Such a pyramid shape is realized most evidently in Bamberg, and minimally so in Utrecht, where the whole configuration appears more of a trapezoid. In sharp contrast, in P11 animals and birds seem to be positioned on the same single plane—rather than along separate multiple lines—which slopes upwards continuously from this side to the back. In other words, the animals are arranged on a two-dimensional basis in P11, whereas they are represented elsewhere strictly in one-dimensional, linear terms. The strict linearity in procession is reinforced by the consistency with which the animal faces are without exception illustrated in profile in Bamberg, Millstatt, and P13. The frieze-based representation of animals looks most fitting the name-giving scene (Bamberg, Millstatt, and San Marco), in which the multiple processions of animals, for the most part, face toward the name-giver to wait their turn to be touched and named. Decontextualized from the Naming scene and relocated in the presence of the beholding God, however, the original pattern needs reconfiguration and recomposition in P11 to be conceptualized as a population of animals living naturally on their own in paradise. Similarly relocated, however, P13 retains vestiges of the earlier representation based on a procession of animals in multiple rows, although the pyramid shape is considerably deformed. While analogous to the three members of the Cotton family at first glance, Utrecht seems to be organized distinctly on deeper grounds. The living creatures no longer form a closed group; rather, they are incorporated into a larger pyramid whose peak is occupied by fire. The third distinguishing feature is one of symmetry, by which is meant the orientation of animals in relation to the entire group rather than characterized on an individual basis, as with the first property treated above. Since the Naming pictures (Bamberg, Millstatt, and San Marco) are regulated by the orientation toward the name-giver, there would have been no room for an overriding rule to intervene and effectuate a contradictory, higher-order directionality. Thus, of central interest here is Utrecht 82v. In it, all

150 

 6 Adam and Eve in Paradise: P10/P11/P13

animals are organized into two groups, the right half and the left half, which are both oriented to the center and thus stand face to face with each other in groups. None of the other pictures under examination here exhibits the remotest resemblance to such group-based consistent organization. That neither P11 nor P13 shows the slightest sign of following the Utrecht pattern seems of paramount significance. Since these Junius pictures are dissociated from the Naming scene, the orientation to the name-giver should have no role to play in determining the arrangement of animals at any level of organization. Moreover, given that the Junius artist exhibits a marked preference for symmetry throughout his practice of reorganization and reconfiguration, it is expected that he should act accordingly in this case as well. More specifically, if P11 had been modeled on Utrecht 82v as suggested above, why did it fail to implement an overall symmetrical arrangement of animals as in Utrecht? In view of the analogously symmetrical placement of Adam and Eve on the same page, it would have been quite natural that the animals be depicted in a similar fashion, resulting in maximal symmetry from top to bottom. The artist’s failure to achieve the symmetrical arrangement of animals in a way comparable to Utrecht 82v therefore may be adduced as evidence against identifying this picture as a critical model for P11 and P13. San Marco and Utrecht depict animals in pairs, while the others picture them in singles. Since the pair-based representation obviously should require extra care and effort in composition, it is naturally characterized as more elaborate than the single-based counterpart. That P11 and P13 do not share this refined mode of representation may argue against postulating a special contact relation between Junius and these two works (or their close cognates) such as would have led the Junius artist to adopt this elaborate style. Of vital importance in this connection is the appearance of Adam and Eve, the human couple, in P11 and P13. Such an independent presence of the human pair undoubtedly would have facilitated an analogical extension of the coupling to other living creatures, insofar as a proper model like Utrecht 82v had been accessible. Furthermore, the exclusive selection of the array of animals as a domain of inspiration would have been perfectly feasible in light of the demonstrable influence by Utrecht in the limited zone of pictorial space in P3 (section 3.2). Despite the optimal convergence of favorable conditions, however, the expected borrowing from Utrecht 82v, surprisingly, did not take place in P11/P13. Given that there are independent cases—albeit few—in which Utrecht presumably would have been instrumental—albeit in an auxiliary way—in shaping pictorial designs in Junius (P3.1 and P3.3; section 3.2), the Junius artist certainly would have had access to the Utrecht Psalter itself. Thus, despite presumable accessibility, the Junius artist opted not to follow the alleged model in this case. This avoidance would in turn suggest that the artist may have been far from enthusiastic about emulating Utrecht, as he obviously missed the opportunity to maximize symmetry in his best interest. For further discussion on the failure of Utrecht 82v to inspire P11/P13, see section 12.6.1 below. It should be noted in passing that the pair-based arrangement does not contradict the requirement for unidirectionality that holds for the Naming pictures, as testified eloquently by San

6.2 A diachronic–comparative perspective 

 151

Marco. Completely independent of the feature of orientation, the refined arrangement may be implemented accordingly.96 Finally at issue are the varying ways in which birds are depicted. There are a few different aspects involved. First, San Marco does not offer a single instance of a bird in the Naming scene, which is obviously at odds with the Genesis text. According to Genesis 1:19–20, Adam named not only animals but also birds. Thus, San Marco must be regarded as peculiar in this respect. This peculiarity, however, is not shared by Junius: in both of the relevant pages, birds are illustrated. The lack of this marked property in Junius, however, is of no heuristic value in shedding light on its specific sources. More specifically, since Junius does not share this peculiarity of excluding birds from depiction, no special derivational relationship between the two—Junius and San Marco— may be justified in this regard. A second aspect has to do with the pose of birds. In Bamberg and P11 the birds are all standing on the ground; no flying or perching figures are in evidence. In the remaining pictures (except for San Marco, which contains no birds), by contrast, birds are depicted flying in the air or perching on the branch. Of particular interest is the recurrent lack of agreement between P11 and Utrecht 82v, both of which, without any intrinsic connection with the Naming scene, should be organized in parallel, particularly given their respective predilection for symmetry. Thus, if the Junius artist had followed Utrecht 82v in composing P11, he would more likely have arranged birds on the branches and depicted them flying, as in P13, which is not borne out by the facts. The representation of birds only walking on the ground in P11—despite the possible exemplar of Utrecht 82v to the contrary—therefore may be attributed to the same manner of depiction in Bamberg, in which the animals and birds are arranged in procession on the ground. Put another way, P11 and P13 would have been inspired—in their differing ways—by Cotton Genesis models—such as exemplified by Bamberg 7v and Millstatt 9r—to the exclusion of Utrecht 82v. To recapitulate the derivational positioning of P11 and P13 in relation to the group of resembling pictures comprising Bamberg 7v, Millstatt 9r, San Marco, and Utrecht 82v, the divergence between P11 and Utrecht 82v must be stressed despite their appearances to the contrary. The absence of a marked preference for left orientation, which P11, P13, and Utrecht 82v share in distinction from the three Cotton Genesis members, simply may be explained as contingencies due to the default neutrality to directionality, which obviates postulating a derivational relationship between Junius 11 and Utrecht. Had there been a substantial contact between the two, P11/P13 and Utrecht 82v would readily have embodied common compositional innovations by virtue of their shared conceptual bases—the dissociation from the Naming scene and the preference for symmetry among others. Actually, however, P11/P13 failed to adopt the overall symmetri-

96 Seen in this light, the sharing of the pair-based representation may lead one to suspect an intimate relationship between San Marco and Utrecht, but details await further investigation.

152 

 6 Adam and Eve in Paradise: P10/P11/P13

cal arrangement and the pair-based grouping of animals distinctively characteristic of Utrecht 82v. Thus, P11/P13 were not significantly influenced by Utrecht 82v. By contrast, a number of differences on the surface between P11/P13 and the members of the Cotton family, notably Bamberg 7v and Millstatt 9r, are subject to genealogical accounts by deriving the former from the latter and invoking well-motivated readjustments consequent of their reorganization as a depiction of Adam and Eve in paradise as opposed to the Naming scene. Of further interest, P11 and P13 slightly differ in their conformity to the original representation of the Naming, as reflected in Bamberg 7v and Millstatt 9r. Having undergone a less thorough recomposition, P13 still retains a frieze-based representation of the animals and a vestige of their overall placement in pyramidal form that are reminiscent of Bamberg 7v and Millstatt 9r. Quite understandably, the larger space reserved for P11 than for P13 would have required a greater amount of planning and resources for composition, which would have further facilitated innovation in design and execution.97 On the whole, then, P11 may have been derived from a model commonly reflected in Bamberg 7v1 (right) and Millstatt 9r, the Naming of the Animals. The three horizontally concatenated zones of Adam, the animals, and God in the models were rearranged along the vertical axis, the animals (below), Adam and Eve (middle), and God (above). The original pyramidal frame exclusively concerned with arrangement of animals was extended, in conjunction with the vertical reorientation at issue, to cover God and the couple, with God standing at the top. While the frieze-based organization of animals was deconstructed, the overall pyramidal frame was inherited intact but reconstituted in an expanded form to embrace all figures. Incidentally, the verticalization implemented may be adduced as a parallel case for the composition of P6/P7, whose major model is attested in less transparent form at Salerno, however (section 4.2). The (partial) indebtedness of P11 to the pictures of the Naming of the Animals, as asserted above, has heuristic implications for a diachronic account of P13. Since the original scenes attested in Bamberg, Millstatt, and San Marco present Adam as the name-giver, and since the adapted version, P11, treats Eve as well as Adam in paradise, an analogical expansion would have been at work in the transition to broaden the range of the referents involved, from Adam in the singular to Adam and Eve in the dual. On the basis of such a referential expansion, then, one may be led to infer that sharing the depiction of the couple and the animals in paradise, P13—a closely associated variant of P11—would have evolved through a comparable expansion from Adam to the couple of Adam and Eve in subjecthood. Actually, there is an independent scene in existence, “Adam in Paradise,” attested as it is in San Marco (Pl. 41) and Millstatt 8r (Pl. 60), two of 97 As argued and speculated in section 5.2 above, the Creation of Adam would have been pictured in one of the lost three leaves (cf. Doane 2013: 11). Given the extensiveness of the missing part, one may reasonably expect that the Naming of the Animals also would have been included there; conceptualized as the Naming scene proper, this illustration presumably would have been more faithful to the Bamberg 7v2 and Millstatt 9r models than P11 by adhering to the strict format of multiple friezes.

6.2 A diachronic–comparative perspective 

 153

the three works under consideration that preserve the Naming scenes (Pl. 42 and Pl. 62, respectively), which appear immediately after the pictures of Adam in Paradise. While a corresponding scene is absent in Bamberg, the Naming scene obviously presupposes such a setting. Thus, one may infer that the presence of Adam in Paradise would have been absorbed into the following phase of closely associated events, that is, the Naming scene, in Bamberg. Interpreted this way, the two scenes at issue are distributed exactly the same way in Bamberg (with a secondary conflation), San Marco, and Millstatt. The conceptual and pictorial inseparability of the two scenes has been thereby demonstrated beyond reasonable doubt. In this light, it should be a foregone conclusion that P13 would have derived largely from the scene of Adam in Paradise through multiplication of a human referent (Adam/ Eve instead of Adam), in parallel to the analogous derivation of P11 from the pair of the Naming scene, that is, Bamberg 7v2/Millstatt 9r. Of particular interest is the striking similarity—virtual identicalness—between P13 and Millstatt 8r (Adam in Paradise) with respect to Adam’s gesture. Of further significance are the two lamps hanging from top in Millstatt 8r. Although no architectural structure is depicted there, these objects may imply such a framing, more specifically, an arched ceiling like the one represented in P13. On these grounds, it may be warranted to make a more substantive hypothesis (than formulated at the beginning of this paragraph) that P13 originated from a model that is reflected in Millstatt 8r most transparently. Still awaiting explanation is the seemingly unique hexagonal frame surrounding God in P11. Three sources of inspiration may be identified as responsible for its formation: (i) the image of Christ in Majesty in the Carolingian, notably Touronian, Bibles and Gospel Books; (ii) the representation of concrete buildings, as in Vivian 386v2 (Pl. 31), San Paolo 3v2 (Pl. 33), St. Aure Gospel (Court School of Charles the Bald; Paris, Bibliothèque de L’Arsenal Ms. 1171), fols. 17v, 72v, 108r, 164v; (iii) the figure of God conceptualized in unity with the seraphim, as uniquely represented in Pii. As regards the first source, the hexagonal frame in P11 seems similar to a lozenge-shaped outer frame, attached with circles at the four edges, surrounding the enthroned Christ in a mandorla, as in Vivian 329v (Pl. 30) and San Paolo 259v (Pl. 36). While the replacement of the normal angles with the circles may make the resultant shape look like an octagon, the underlying rhombus—as manifested most transparently in Grandval 352v (Pl. 25)—may resist such a full recategorization. Furthermore, the four circles may receive varying degrees of perceptual prominence depending on context. Specifically, the vertical and horizontal axes may affect their perceivability differentially, augmenting it in one direction and diminishing it in the other. Given that P11 is reorganized primarily along the vertical axis, as argued above, the top and bottom circles may have come to figure more prominently than those on the sides. This prioritization of the vertical over the horizontal then would have given rise to the recategorization of the outer frame as hexagonal in its overall shape. Of no less interest are the two scribes in the lower right and left (Vivian 329v, Pl. 30; San Paolo 259v, Pl. 36), who may be compared with the figures of Adam and Eve looking up in P11. Thus, the configuration of the two evangelists looking up at the

154 

 6 Adam and Eve in Paradise: P10/P11/P13

figure of Christ at the center enclosed in a lozenge constitutes a compositional parallel to and an analogical model for P11. While the analogy with the image of Christ in Majesty in some Touronian Bibles provides an abstract scheme and an inspirational impetus for the hexagonal frame in P11, the concrete form—a crenellated city wall—in which it is eventually embodied would have been modeled on substantive architectural structures. At issue are comparable buildings such as are depicted in Vivian 386v2 (right) and San Paolo 3v2 (left). As pointed out by Broderick (1978: 184), the Gospel Book of St. Aure 17v, 72v, 108r, and 164v also may be relevant. Of overriding importance here is the walled city in Vivian 386v2. Only the front side is provided with the gate. Significantly enough, the enclosure in P11 has a noncrenellated portion exactly on the corresponding edge—suggestive of the free passage—facing Adam. Thus, a concrete image comes from these city walls, while a more abstract conceptual scheme is based on the motif of Christ in Majesty. As revealed above, the three-part configuration—the framed and enthroned figure of Christ flanked on the lower sides by the evangelists—provided a compositional scheme for P11. There is another dimension to this underlying structure. When one focuses on the circular contours enclosing the three figures—God in the center and the two evangelists in the lower corners—to the exclusion of the figures themselves and the rhombic outer frame, one obtains two arcs and a circle or ellipse at the joint of these flanking arcs. The emergent structure proves to be well-matched with the upper architectural frame—the double arches on both sides and the mandorla in the center at their joint—of P13 in its overall contour. Thus, the upper frame consisting of the double arches and the mandorla in P13 most plausibly may be derived from the image of Christ in Majesty in Vivian 329v and San Paolo 259v. Given the synchronically motivated derivational relationship—the variation on the invariant motif—between P11 and P13, the structural link of P13 to the image of Christ in Majesty in these works reinforces the latter’s association with P11. As substantiated in the previous section, P11 and Pii are situated in parallel in relation to their respective variations, P10 and P2. Both are characterized as elaborate representatives of the groups they head, commonly providing a frontal, symmetrical depiction of God in Majesty. Given such structural and conceptual parallelism, the one image resoundingly evokes the other. Accordingly, Pii would have been implicated in no small measure in guiding the composition of P11. Specifically, since the figuration of God is a common concern, while the representation of animals exclusively bears on P11, the hexagonal contour of the seraphim surrounding the figure of God and also the angled arch resembling the upper half of a hexagon in Pii more likely than not would have promoted use of the hexagonal enclosure in P11. In conclusion, the appearance of the hexagonal frame would have been determined by an interplay of the formative forces that differ in levels of abstraction in which they operated and in the sources from which they originate—namely, internal versus external to Junius 11. At the deepest layer is the resounding echo of Pii, which conjures up a hexagonal aura around the figure of God. Next higher up along the scale is the Tou-

6.2 A diachronic–comparative perspective 

 155

ronian visualization of Christ in Majesty that is embodied in Vivian and San Paulo, in which Christ is framed in a lozenge with circles at the four corners toward whom the two evangelists look up from below. Finally, at the most concrete level of materialization, a hexagonal city wall with a crenellation is accessible for adaptation and reworking in the same works referred to in connection with the second layer of influence.

7 The Temptation of Adam: P20/P24 7.1 A synchronic–structural perspective P20 (Pl. 12) is divided into two fields of representation, paradise in the upper half and hell in the lower. In between is found a passage from one to the other, the gate of hell. Continuing from the last picture (P17, Pl. 11) and largely overlapping it, the hell scene of P20 visualizes afflictions Satan and his company were doomed to suffer in hell. The illustration corresponds to the text running from approximately Genesis B 313a to 441a (incomplete) in the extant manuscript,98 in which Satan laments his fate, voices his strongest enmity against God and mankind, and calls for his followers to deceive man into breaching God’s commandments. In the interface between the two worlds, one of the devils passes through the gate to carry out Satan’s order as his agent. This visualization thus corresponds to Genesis B 442a through 453b. As Ohlgren (1972a: 204; 1992: 91; see also Ohlgren 1991: 15–16) describes, the devil standing at the left corner facing Satan and gripping his hands may well be identified as the messenger who is shown at the gate departing for paradise. In this interpretation, the bottom scene of Satan and the devil standing in front of him indicates that the devil in question is being commissioned to exact revenge on humans on Satan’s behalf (cf. Gollancz 1927: xli; Ohlgren 1992: 91). The striking similarity of the two figures—Satan and his agent—in appearance and size, as observed in Gollancz (1927: xli), would have been designed as a visual representation of the messenger’s wholesale fidelity to, and hence his virtual unification with, Satan, in accordance with the textual description. While the hell picture thus yields a straightforward reading, its paradise counterpart poses difficulties of interpretation. Obviously, however, the upper half consists of two separate scenes. On the right side, Adam and Eve, largely superimposed over one another, turn rightward, pointing to the tree standing to their right. Correspondingly, this may be regarded as a visualization of the verses from 454a through 490b, beginning with the following passage (Genesis B 454a–460b), which fits neatly the pictorial representation: oððæt hē adam / on eorðrīce, // godes handgesceaft, / gearone funde, // wīslīce geworht, / and his wīf somed, // frēo fægroste, / swā hīe fela cuðon // godes gegearwigean, / þā him tō gingran self // metod mancynnes / mearcode selfa // and him bī twēgin / bēamas stōdon (until he [= Satan’s messenger] found Adam, God’s creation wisely wrought, ready in the earthly kingdom, together with his woman, the fairest lady, as they knew how to bring about much good, whom the creator of mankind himself designated as his underlings. And they stood between two trees…) (Anlezark’s [2011: 37] translation)

98 There are two leaves missing between 441a (end of page 22) and 442a (beginning of page 23; Table 2, section 1.1); see Doane (2013: 5, 11–12). https://doi.org/10.1515/9783110788068-007

158 

 7 The Temptation of Adam: P20/P24

Thus, the upper right half of P20 represents Adam and Eve’s life in paradise, as perceived through the eyes of Satan’s agent upon arrival from hell. In the left scene, the serpent winds around the tree and confronts Adam (originally, despite the apparent identity as Eve due to subsequent alterations to the original identity; for details, see below) at close range, who glances back to the serpent, as described particularly in verses 491a–494b: wearp hine þā on wyrmes līc  / and wand him þā ymbūtan  // þone dēaðes bēam  / þurh dēofles cræft. // genām þǣr þæs ofætes / and wende hine eft þanon // þǣr he wiste handgeweorc / heofoncyninges. (He cast himself then in the likeness of a serpent and wound himself about the tree of death through his devil’s skill, took some fruit there and turned back to where he knew the king of heaven’s handiwork to be.) (Anlezark’s [2011: 39] translation)

In light of the distinct textual alignments in terms of narrative development, the two scenes in paradise should be characterized as sequenced from right to left. Thus, the narrative sequence in the upper half of P20 is from right to left, as correctly suggested— although no argument offered—by Gollancz (1927: xli) as an alternative interpretation of the ordinary view from left to right. This demonstrable temporal ordering apparently conflicts with what the viewer would otherwise be induced to expect: assuming the movement to be continuous in time without disruption and mapping what appears to be a self-evident, vertical movement of Satan’s messenger from bottom to top at the far left of the page, one may be led in anticipation to interpret the right-side scene in paradise as the last of the four–frame pictures on this page, as succinctly characterized in Karkov (2001: 11; see also Karkov 2011: 237): “The artist has arranged the figures to lead our eyes in a circular motion from the figure of Satan bound in hell, to his servant passing through the gates of hell, to the serpent and Eve in the upper left part of the drawing and on to the figures of Adam and Eve in the upper right.” However natural this progression of events may seem at first glance, it proves to be no more than a conditioned expectation that is contingently favored by a clustering of perceptual preferences, rather than derived from general principles as inevitable consequences. First, the left-to-right order may be favored over the reverse direction, influenced in no small measure by the corresponding writing direction. Second, the facial orientation of the couple on the right, in conjunction with their pointing gesture, may suggest a movement in that direction, namely from left to right, unless indications to the contrary are overriding. Third, a single consistent movement—“a circular motion” as Karkov (2001: 11) put it—may be easier to process than complex movements going in opposite directions. In other words, unidirectionality is optimal, at the expense of bi-/multidirectionality. These expectations, however, are matters of relative plausibility rather than of absolute necessity. In the presence of evidence to the contrary, one is required to reject them in favor of alternatives that turn out to be more harmonious with the available indications.99 99 Having succumbed to the above biased left-to-right conceptualization, Raw (1953: 49) speculated that the two scenes at issue in P20 may have been reversed by misplacement. Originally, the pair of Adam and

7.1 A synchronic–structural perspective 

 159

The right-to-left sequence has been substantiated on intrinsic grounds to be in function in other places of Junius 11. In P3.1a, the attending angels are proceeding from right to left, contrary to the model images in Vivian and San Paolo (section 3.1). Similarly, the two constituent scenes of the Creation of Eve (P9) develop from right to left, again in distinction from their models in Vivian 10v (section 5.1). In addition to having its own feasibility enhanced by parallel representations elsewhere, the right-to-left sequentiality in the upper section of P20 is predicated on the textual progression with which it is pictorially aligned, as argued above. By far more convoluted than the issue of directionality is the identity of the figure that the serpent is addressing in the upper left of P20. According to the conventional view (Gollancz 1927: xli; Raw 1976: 140; Ohlgren 1992: 91; Gameson 1995: 44; Ericksen 2001: 49; Karkov 2001: 11), the figure in question must be Eve, as indicated in what appear to be unequivocal ways by its physical attributes, namely, the nipples and the long hair, which are in marked contrast to their corresponding features of the man, Adam, standing to the right of the central tree and situated in front closer to the viewer (cf. Broderick 1978: 200). Thus, these few physical properties apparently substantiate the identification of the central figure in P20 as Eve, as far as they are considered on their own from a strictly local perspective. However, by widening perspectives and relating the above minimal traits to the overall depiction of the figure at issue and further to the representation of the couple in the right scene, one may realize that the conventional identification is open to criticism and far from unequivocal. In fact, despite such seemingly compelling indications to the contrary, there are overriding reasons for identifying the figure in question as originally intended to be Adam, although this original characterization is no longer transparently reflected in the extant copy. First, the depiction of the figure itself seems at variance with the biblical narrative of the Temptation of Eve and the corresponding conventional pictorial representation as manifested in Grandval 5v3 (Pl. 24) and Vivian 10v2 (Pl. 27), in which the female figure confronts the serpent squarely and trustfully. Instead, in P20 the figure fails to exhibit its wholesale submission to the snake by its turned-back posture and its restraining hand gesture. Second, the upper right scene of P20, commonly regarded as a sequel to the left image (see above), would cast doubt on the identification of the immediately preceding scene as the Temptation of Eve. While the Temptation of Eve resulted in the Fall of Adam and Eve in the Book of Genesis 3, the ways in which the couple is depicted in P20 (right) seem incongruent with the Fall. Specifically, the aversion of both Adam and Eve from the snake and/or the central tree— in their largely overlapped posture, which suggests their determined unity in obeying God’s order—would show persuasively that they are not yet fallen, but remain true to God. Of interest in this connection are the figures of Adam and Eve in P10, who are

Eve, occurring at left, would have corresponded to a state before the temptation of Eve depicted at right, in greater harmony with the narrative content. For a criticism of Raw’s view, see the following section.

160 

 7 The Temptation of Adam: P20/P24

shown fully obedient to God. This representation seems very similar to the configuration of Adam and Eve in the upper right of P20. This configurational resemblance may further support the reading that the couple in the upper right of P20 is committed fully to God’s commandments. Given the traditional conceptualization of left-to-right narrative movement in the upper half of P20, then, the conventional identification of the human figure in question as Eve is contradicted by the independently motivated plausible interpretation of the right-half scene. Yet, in light of the new finding established above that the pictorial narrative in the upper zone of P20 advances from right to left, the identity problem under examination might require reevaluation. Insofar as the couple’s obedience to God is represented in the earlier scene (right), identifying the left figure as Eve in the following picture (left) can no longer be rejected for reasons of narrative incoherence.100 Nonetheless, irrespective of the narrative direction involved, the compositional incongruence of the figure’s posture and gesture in the upper left of P20, as adduced as the first point of objection in the preceding paragraph, still militates against feasibility of the traditional identification under criticism, to the advantage of the alternative identification proposed here as Adam. Restated from a contrary point of view, even if the left figure facing the snake had been (re)interpreted to be Eve in the extant manuscript, the corresponding pictorial substantiation executed in support would have been only marginal in scope, bearing solely on the upper body of the figure, and hardly being orchestrated further afield on the configurational and compositional dimensions. This lack of coherence and systematicity in presenting the figure as Eve may lead to the suspicion that this figural identity would have arisen secondarily, contingently, and locally without careful design and systematic coordination. Specifically, given the distinct posture of the figure including the hand gesture and the direction of the legs, it seems hard to believe that the original artist determinedly would have conceived it as Eve in full conformity with the Bible at the expense of Genesis B (or more precisely the Old Saxon Genesis underlying the Old English rendition). Rather, the figure would originally have been intended as Adam in full harmony with the vernacular text, and only subsequently reidentified as Eve. Whether such an alteration was due to an erroneous drawing or an attempt at “improvement” by a later hand, such a pictorial intervention, at whichever stage of transmission it may have taken place, would have been minimal and superficial in its scope of effects. The original intension to represent Adam responding to the snake remains in force despite the partial cancellation subsequently executed to override it. 100 Accordingly, inasmuch as one adhered to the innovative right-to-left narration introduced by the original illustrator, one would have encountered no hinderance in restoring the conventionally feasible identity as Eve for the left figure in P20 in accordance with the cognate pictures in the Cotton Genesis members and in defiance of the Touronian innovator responsible for illustrating the Old Saxon Genesis, the Touronian antecedent of Genesis B. For details on the reinterpretation as Eve resulting in the current image of P20 (upper left), see below.

7.1 A synchronic–structural perspective 

 161

In addition to the above argument explored on internal grounds, an external circumstantial reasoning may be brought to bear on the present issue for reinforcement and reconfirmation. If the figure on the left side of P20 had been Eve in the original, it would follow that no illustration would have been provided for the unsuccessful temptation of Adam that is treated specifically and dramatically in Genesis B 495a–546b. Given the importance of this unique episode as reflected in the extensive passage devoted to it, it would be hardly intelligible why it failed to receive pictorial representation in comparable fashion.101 Thus, the otherwise complete absence of visualization for the Temptation of Adam, despite its outstanding thematic status in the vernacular text, may be adduced as a supplementary argument against interpreting the figure to have been designed as Eve in the original. A more theory-dependent justification may be adduced further to make the originally intended identity as Adam more plausible. The identification with Adam in P20, paired with Satan’s agent, would constitute a maximal contrast to the pair of God and Eve in the upper left zone of P9: in the corresponding panel of P20 Adam, standing higher and close to the center, looks back to Satan’s messenger to the left, with one of his hands holding the other with a restraining gesture; in contrast, located at the center and in a higher position, God faces Eve frontally, with one of his hands touching Eve’s. As argued at length in section 5.1 above, P9 and P20 are fully contrasted in almost all details, with God on the one hand and Adam on the other figuring close to the center (in spatial terms) and dominating the scenes in paradise (in narrative terms), not only in the left but also in the right zones of P9 and P20, respectively. In view of these formal and thematic divergences and parallelisms, the identification of the left human as Adam in P20 would create a perfect contrast with P9 in terms of composition and configuration. Subscribing to the conventional identification as Eve and addressing different concerns, Karkov (2007: 62) points out a structural correspondence between the figure of Eve just created by God (P9) and that of Eve being spoken to by Satan’s messenger (P20): “In the upper left corner of the page [= P20: S. S.] Eve gazes into the mouth of the serpent, as she gazed into the face of the Lord at the moment of her creation.” The alleged correspondence, however, is only a matter of a secondary and accidental nature, derived as it is from a subsequent interference, whether an “error” or “correction.” It should be noted that, referring to “Eve’s” hand posture in the sentence that follows the above quotation, Karkov simply mentions its pointing gesture while disregarding its abstaining pose, as traditionally acknowledged. Doing full justice to the figure’s gesture and its implications (see below), however, would call into question Karkov’s (2007: 62; see also Karkov 2001: 11) interpretation of the paradise scene in P20, which favors the biblical reading at the cost of the Genesis B rendition: “The illustration provides a visual summary of the narrative action of the fall in which the serpent will deceive Eve who 101 It should be noted that, according to the usual interpretation, P24 also treats the Temptation of Eve. For a criticism of Karkov’s (2001; 2007) distinct interpretation that P24 depicts Adam being tempted, see below.

162 

 7 The Temptation of Adam: P20/P24

will subsequently deceive Adam, . . .” After all, there seem to be no independent grounds for substantiating Karkov’s stipulation, particularly for inferring that the picture would have been designed to be in line with the Bible, even in apparent contradiction to what is being narrated in the text that the picture accompanies. In light of the foregoing arguments, it is unlikely that the identity of the figure as female (Eve) would have been conceived of carefully and firmly, and accordingly executed thoroughly and consistently, in the original visualization. The remaining representation in the upper half of P20, which is largely at variance with the corresponding biblical text, would instead testify that the apparent identification as Eve was far from what was originally intended; rather, it was only derivatively introduced due to a contingent error or as a result of superficial revision in the course of copying. The suggested alteration from Adam to Eve in the identity of the upper left figure obviously requires further consideration. Of particular interest in this regard are the identical attributes of the left figures in P20 and P24 (Pl. 13): the long hair, the nipples, and, most significantly, the restraining gesture; compare Broderick (1978: 206).102 These shared features would suggest that the artist would hardly have made mistakes in figuring them as Eve independently on both occasions. Of relevance in this connection is Broderick’s (1978: 203–204) view that the artist, in the course of drawing P20, became aware of the misrepresentation he had just executed of the left figure as Eve, and thereupon made necessary adjustments to minimize adverse effects. Carrying this supposition to its logical conclusion, the figure in P24 more likely would have been unaffected by the mistake introduced in P20 and hence represented as Adam free of error: by the time the artist turned to engage with P24, he would have completed his modifications and resumed his work with due awareness of the gender identity of the figure. Thus, the persistence with which the error was made defies explanation in Broderick’s view. Moreover, according to Broderick (1978: 204), the chief adjustments that the artist implemented for improvement upon detecting his error was the addition of Eve behind Adam in the right picture (P20): this alteration, Broderick claims, was meant to suggest that Adam had been tempted earlier than Eve as depicted in the left part. It remains unclear how such an image manipulation would have served the purpose of removing the conceptual discrepancy just arisen between textual and pictorial representation. In more general terms, there is cogent reason for calling into question Broderick’s interpretive framework itself, given the right-to-left sequence in the upper portion of P20: the artist more likely would have illustrated the right half first and then moved onto the left. This would lead to the inference that, after finishing the upper right picture in P20 as it is, the artist—intentionally or accidentally—(mis)represented the left figure 102 The (almost) identical representation of the left figures in P20 and P24 seems to indicate that the artist would have intended them to be one and the same human, Adam or Eve; it is unlikely that he would have designed them to be different in their identity, Eve (P20) and Adam (P24), in contradiction to Karkov’s (2001: 11–14; 2007: 62–63) seemingly ingenuous claim. For a further criticism of Karkov’s proposal, see Broderick (2009: 391).

7.1 A synchronic–structural perspective 

 163

as Eve, presumably due to the conventional force of the familiar image of Eve being tempted by the serpent, as assumed by Broderick himself (1978: 204–205), as well as Raw (1976: 141) and others. The ultimate cause for the misrepresentation thus must have been one and the same, regardless of disagreements in the finer details. Yet, according to the alternative conceptualization being explored here, the gender alteration introduced secondarily would not have resulted in compositional readjustments for the right picture, but remained strictly localized in its effects bearing solely on the upper body of the figure in question. Apparently without becoming aware of the resultant conceptual incongruence between the appearance of Eve at left and the couple of Adam and Eve at right, the copyist/artist continued to allow himself to be trapped by the biased identification and made the identical error (or correction; see below) for a second time in P24. Thus, the persistent change in the figure’s identity may be better understood by invoking the determining power of the traditional biblical image without postulating the artist’s (dubious) attempt to minimize the error. Perhaps the copyist/artist would have remained so much bound by the biblical narration and the corresponding visual representation that he was totally unaware of the mistake he had committed. Or rather more likely, he would deliberately have changed the identity from Adam to Eve “for pictorial improvement” in line with the familiar biblical tradition. It should be noted in this regard that, either way, the resultant revisualization of the Temptation of Adam as that of Eve in accordance with the biblical story would not immediately bring about a contradiction with the narrative content of Genesis B. The reinterpretation would simply delete visual references to the Temptation of Adam altogether and instead address directly the Temptation of Eve, at a point much sooner than was originally designed (P28). Thus, this erasure of the Temptation of Adam unlikely would have incurred outrageous inconsistencies with the Genesis B text. On the contrary, in pictorial terms it would have masked—rather than exposed—the narrative difference between the Book of Genesis and the Old English/Old Saxon Genesis. The prospect of the ensuing lack of manifest contradictions with the vernacular Genesis may have facilitated replacing the Temptation of Adam with that of Eve. One should be reminded further in this connection that P24, the illustration that appears next to P20 in the current manuscript, presumably would have other pictures preceding it and subsequent to P20 because of the two missing leaves between 441a (disrupted) and 442a, as mentioned above (Doane 2013: 5, 11–12; see Table 2, section 1.1 above). In view of such a long distance between P20 and P24 interspersed by other drawings, it seems even more unlikely that the artist would make exactly the same mistake of representing the left figure as Eve (at least) twice independently and at the expense of the model pictures at hand, unless induced persistently and irresistibly by the biblical-pictorial imagery of Eve being tempted by the serpent.103

103 The missing leaves may have contained a drawing or two, which constitute variants of P24, as with the series of related pictures comprising P10, P11, and P13.

164 

 7 The Temptation of Adam: P20/P24

That the artist readily succumbed to the biblical image of the Temptation of Eve— however widespread in currency and compelling in force—in defiance of the vernacular narrative to the contrary may in turn make it natural to assume that the man responsible for the persistent misrepresentation would likely have been far removed from the original situation of manuscript illumination in which the illustrator had access to the details of the text and was directly exposed to the formative forces of manuscript decoration. In distanced working environments of that kind, his work would have tended to be denied effective monitoring, guidance, and feedback. More specifically, the pictorial distortion at issue presumably would have been added in the course of transmission well after the initial production, that is, in the derivative circumstances in which the original intentions and earlier shaping factors were no longer in full function or reliably recoverable. Now that the sequence from right to left on the one hand, and the identity of the figure facing the serpent as Adam on the other, have been established beyond reasonable doubt, one may summarize the ongoing pictorial narration in the upper half of P20. Beginning with the right side, the earlier of the two scenes in paradise, Adam and Eve are found standing between two trees (Genesis B 460a–460b), the Tree of Life (at the right) and the Tree of Death (at the center), as described in detail in Genesis B 467a– 485b. By almost superimposing one figure on the other in representation, the artist seems to insinuate that the couple are united completely in obeying God: in unison, they point to the tree to the right and conversely turn away from the one on the left. Such full conformity to God’s commandments, as depicted earlier in P10, P11, and P13, is echoed most eloquently by a partial replication of the P10 drawing in P20. Of particular interest is a similarity in the way the pair is configured, as observed above. While the two figures in P10 do not overlap, to be sure, the overall resemblance between the two pairs appears to be striking enough to evoke one other. This resemblance, moreover, might go so far as to hint at an implied presence of God to the farther right beyond the margin of the page (to be elaborated in diachronic–genealogical terms in the following section). At the same time, the opposition between the upper and lower bodies in terms of their orientation relative to the tree at the center—the Tree of Death—may be taken as an indication that Adam and Eve are situated in an ambivalent position, subject to two antagonistic forces beyond their control. In fact, encountering the couple’s absolute submission to God when arriving in paradise, Satan’s messenger attempts to deceive Adam, as illustrated in the left side and verbally described, starting at Genesis B 491a–491b. Adam’s restraining hand gesture, as conventionally conceptualized, shows that despite the messenger’s cunning, Adam is determined not to be deceived, with the right hand guarding the left from picking fruit from the left tree—the Tree of Death—entwined by the snake. Thus, the order of events in P20 is such that the vertical axis from hell to paradise and the horizontal axis from right to left converge at the upper left zone in culmination, where Satan’s agent encounters God’s human creature (Adam) face to face in irreconcilable opposition.

7.1 A synchronic–structural perspective 

 165

Corresponding to the multiplicity of the episodes in space and time illustrated on the whole page, P20 is implicated in a network of interrelations with a variety of other pictures in Junius 11, syntagmatically concatenated or paradigmatically associated and contrasted. Accordingly, at this point it may be useful to recapitulate the pictorial linkage in force. Considered as a whole, the page constitutes the polar opposite of P9, as discussed in detail in section 5.1 above. In their fundamentals, P20 illustrates the appointment of one of the devils as Satan’s emissary, whereas P9 visualizes the creation of Eve. Both—along with Adam—are proxies of Satan and God in paradise. These two acts thus may be regarded as equivalent in analogical terms. In this light, it becomes more amenable to understanding why Satan’s emissary is transformed in the similitude of Satan himself, as noted in Gollancz (1927: xli) and elaborated above. Just as Eve, like Adam, is created in the likeness of God (Gen. 1:26) and by implication of angels,104 Satan’s corresponding act in hell—the commissioning of his messenger—reshapes one of his followers thus chosen to resemble himself in appearance. Upon arrival in paradise, Satan’s agent presents himself to Adam as God’s messenger (Genesis B 497b–499a; see below). This angelic disguise, in form and function alike, puts into relief the underlying irresolvable opposition in the messengers’ identities that is powerfully symbolized by the two contrastive passage scenes from hell to earth on the one hand (P20) and from heaven to earth on the other (P9). Concerned in general terms with the torment and torture of Satan and his cohorts in hell, the lower part of P20 constitutes a variation to that of P17. Two points of contrast, however, are critically involved in forming the unique identity of P20 in distinction from P17. First, while Satan’s abode is sealed off in complete enclosure in P17, it has an opening to paradise in P20, which makes Satan’s machinations possible to implement. Second, in P20 the gesture of touching Satan’s hands singles out one of the devils as of primary importance in determining the fate of mankind. Satan’s call for taking vengeance on God and mankind, depicted in the lowest part of P20, ends in his agent’s report on the successful completion of his mission in the corresponding place in P36 (Pl. 17). The representation of the messenger’s earlier departure on the mission in the bordering zone of hell and paradise in the middle of P20 finds its returning counterpart in the center of P36. The right-side scene in paradise of Adam and Eve’s strict observance of God’s commandment constitutes a sequel to God’s Admonition/Blessing represented in P10 and P11, and then to their subsequent living on earth, true to God, in P13. Thus, standing between the two tall trees, which look similar to those in P13, the drawing at issue in P20 may be characterized as a close variant of P13, both representing the couple’s loyalty to the Creator. In regard to the pair’s harmonious right-oriented posture, however, this picture also exhibits a close affinity to P10. The image at issue may signify Adam and

104 It may be recalled here that specifically at Genesis A 185b the couple is compared to angels in their beauty.

166 

 7 The Temptation of Adam: P20/P24

Eve’s uniform fidelity to God; they are shown behaving, without discord between them, in accordance with God’s commandment. Yet the partially overlapped figures of Adam and Eve in P20, in distinction from their discrete individual figurations in P10, may militate against fully subsuming the two pictures under a single identical type. To generalize, the four pictures under discussion are composed in a complementary manner along three parameters. First, the depiction of the couple standing between the trees puts P11, P13, and P20 together. Second, the uniform orientation of the pair is commonly materialized in P10 and P20. Third, the presence versus absence of God in paradise distinguishes between P10/P11 and P13/P20.105 In this way, P20 articulates with the three previous pictures at distinct junctions. Overall, then, P10, P11, P13, and P20 (the upper right corner) constitute variant realizations of the same underlying scene-type, exhibiting among them varying degrees of family resemblance. The temptation of Adam by Satan’s messenger, as depicted in the upper left part of P20, is represented further in a variant form in a continuing picture, P24. Corresponding to Genesis B 495a–546b, the Temptation here is illustrated from a different perspective in which the devil sent from hell assumes an angel’s appearance, in accordance with the following passages: Genesis B 497b–499a “ic eom on his ǣrende hider // feorran gefēred, / ne þæt nū fyrn ne wæs // þæt ic wið hine sylfne sæt” (I have com traveling here from afar on his [= God’s] errand, it is not long now since I was seated beside him; Anlezark’s [2011: 39] translation); and Genesis B 538a–539b “þū gelīc ne bist // ǣnegum his engla / þe ic ǣr gesēah” (You are not like any of his angels who I saw before; Anlezark’s [2011: 43] translation). While the horizontal arrangement of Adam and the messenger is inverted, their vertical relative positioning remains consistent: in P24 Adam is situated in a higher position than Satan’s emissary, precisely as in P20. Presumably, the scene pictured in P24 is from the opposite side of the viewer’s standpoint as in P20. The difference between the two pictures accordingly may be characterized as primarily consequent to the two complementary perspectives from which to observe the same episode, rather than two fully separate events, as claimed, for example, by Karkov (2001: 11–14; 2007: 62–63). The common identity of the two human beings at issue has its plausibility enhanced by Adam’s identical restraining gesture, to the detriment of Karkov’s view. No less telling than the hand posture is the presence of a short plant between the two large trees. While in P20 Adam is located to the right of this small plant, he stands to the left of the corresponding one in P24. Although the two plants are not illustrated exactly the same way, the commonality in size and location seems to be plausible enough to identify them with each other. The reversal in the positioning of the human figure and the plant accordingly may be explainable as a consequence of the change in perspective, as proposed here.

105 Furthermore, P13 and P20 are in turn differentiated by the varying extent of reduction to which God is represented. P20 constitutes a maximally reduced variant in which God is depicted in zero form. Implications of this analysis, both synchronic (section 12.2) and diachronic (section 7.2), will be explored below.

7.2 A diachronic–comparative perspective 

 167

7.2 A diachronic–comparative perspective The original artist engaged for the first time to illustrate (an Old Saxon antecedent of) the Old English/Old Saxon Genesis was challenged to make an appropriate and adaptive use of the stock of inherited biblical images. He was required not only to represent entirely novel scenes unknown to the biblical tradition, but also to make alterations of traditional sources in accordance with the vernacular narrative to be pictorially accompanied. The upper left scene of P20 is concerned with the messenger speaking to Adam, that is, the serpent’s Temptation of Adam (corresponding to Eve in other cognates such as San Marco; see below); the Temptation scene is thus differentiated from those of Picking and Eating the Fruit, as at San Marco, where the three phases are each represented distinct on their own (Pls. 45 and 46; cf. Kessler 1977: 18–19). The strict separation in Junius is only natural, since, in contrast to Eve, Adam determinedly resists temptation to eat the fruit in the Old English/Old Saxon Genesis. In the Vivian, Grandval, and Bamberg Bibles, however, the first two scenes—Temptation and Taking the Fruit—are conflated into one, so that Eve is being tempted by the serpent, and at the same time she is receiving the fruit from it. While in Grandval the last phase of Eve Eating the Fruit (together with Adam) is treated on its own (Pl. 24), it is merged with the first two in Vivian (Pl. 27) and Bamberg (Pl. 23). In evolutionary terms, then, the upper half of P20 as a whole would appear closest to the Serpent’s Temptation of Eve at San Marco (Pl. 45) in which Adam turns away from Eve interacting with the serpent. In this connection, Raw (1976: 141) refers to the Leaf of the Areobindus Diptych (reverse, top; Hubert/Porcher/Volbach 1970: pl.  219) as a conspicuous parallel to the upper left of P20. Particularly striking is Eve’s posture: she glances back toward the serpent, in contrast to her related representations in which she confronts the serpent, with her right hand pointing to its mouth. As Broderick (1978: 202–203) remarks, however, there remain a few notable differences that should not be ignored, such as Eve not holding her left hand in her right in the diptych. Yet what seems to be a prime distinction—left unmentioned by Broderick—concerns the orientation of Adam, who looks to Eve in the diptych, whereas he turns his back on her at San Marco, exactly as in P20, with the crucial difference in evidence that Eve as well as Adam are depicted to the right of Adam (originally, or Eve, derivatively; section 7.1) in P20, as treated imme­­diately below. According to the narrative progression unique to the Old English/Old Saxon Genesis, Satan’s messenger tempted Adam first. The Junius artist therefore would have been required to change the identity of the tempted from Eve in the original to Adam in the vernacular adaptation. In contrast to the Temptation of Eve in the Book of Genesis, Adam resolutely resisted the temptation, the radical difference that would have driven the artist to make a distinct visualization in some way or other. One may assume here (for the time being and for the sake of argument, though to be rejected eventually) that the San Marco Temptation of Eve (Pl. 45) is closest to the source of P20 in the strongest

168 

 7 The Temptation of Adam: P20/P24

terms, namely, that the work was based solely on the model representation faithfully reflected at San Marco. Then, with the necessary replacement of Eve with Adam, the right figure facing rightward—Adam in the original—would now stand to the right of himself being tempted. Placing the two figures of Adam side by side facing in opposite directions would represent the devil’s failed act of temptation effectively and with minimal alterations added to the model on the surface. In actuality, however, such an image was not brought into being. This is understandable in light of the pictorial organization inherent in the original model. The Temptation of Eve at San Marco (and in its hypothetical antecedent as well) is limited to one scene: Eve being tempted while Adam stays out of it. Through a duplication of the figure of Adam, however, the left and right halves must be conceptualized as separate scenes, temporally occurring one after the other. Thus, the resultant picture is necessarily expanded to a visualization of two distinct events/states, thereby exceeding the capacity of the model. Switching the identities of Eve/Adam to Adam/Eve, rather than duplicating Adam (i.e., Adam/Adam) just given, would have been a minimal alternative adjustment conforming to the limit of one scene imposed by the model. In narrative terms, however, this option would hardly have made sense in the novel context supplied by the Old English/Old Saxon Genesis. A solution was eventually executed, still remaining strictly based on the Temptation of Eve at San Marco, whereby the figure of Adam turning his back on the snake would have been reused as a target of the snake’s temptation. This operation would have resulted in displacing the original figure of Eve depicted in this place. Since Adam is addressed by the snake, however, he is naturally shown turning toward it, but only his face, with the rest of the body oriented in the opposite direction, which is to signify his determined resistance to the temptation. The resultant posture of Adam, then, may be characterized as a composite of Adam and Eve at San Marco (and in its antecedent), the head copying Eve’s and the remaining body carrying over from the turning-away figure of Adam. This is in essence how the upper left half of P20, the Temptation of Adam, would have been derived from the Temptation of Eve in the model as reflected at San Marco. As should be clear, the recomposition so far executed can be accounted for on the sole basis of the San Marco scene in question. The exclusive reference to this picture, however, is incapable of dealing with the right half of P20 in itself, to say nothing of the particular way in which it is filled with the couple largely superimposed and facing rightward. More specifically, the postulation of the Temptation of Eve at San Marco as underlying the upper half of P20 would be incompatible with the appearance of the couple to the right of the tempted. Insofar as one strictly adhered to this San Marco image, one would expect a single figure at most (two options, neither of which was actualized; see the preceding paragraph) or none at all (through a unification of the couple on the left side, thereby leaving no human figure on the right—the option that was allegedly activated; see above). In either case, there would have been no conceptual room for depicting the couple, which requires a separate reference to another model in addition to the antecedent of the San Marco Temptation of Eve.

7.2 A diachronic–comparative perspective 

 169

At this point, one may well invoke Grandval 5v3 (left) as a complementary source. Being an integration of the three scenes that are represented in two separate pictures at San Marco—the Temptation of Eve (Pl. 45) and Eve Picking the Fruit and Giving It to Adam (Pl. 46)—the Grandval picture depicts Eve at the left and the couple at the right within a single panel enclosed by the trees on both sides. In this respect, the two pictures (the Grandval and San Marco cognates), each delineated by panel dividers, would be taken to be equivalent segments to be treated as commensurate, with the snake entwining the tree at the far left as a common denominator. Accommodating two consecutive events side by side, the Grandval variant contains the appearance of the couple on the right side as an indispensable pictorial component, in contrast to the San Marco Temptation, which depicts only a single scene and is hardly open to an augmentation with another scene within its own sphere of visualization, a kind of practice that was executed in the upper half of P20. In this way, complementary to the San Marco Temptation, its Grandval counterpart integrated with the Fall would have been implicated in the designing of P20. Thus, to the right of the tempted Adam, Adam and Eve are available for depiction in a separate but adjacent scene. The artist was then challenged by the following issue: What should he do with this extra scene to connect it meaningfully with the Temptation of Adam to the left? Needless to say, it would be out of the question to mechanically copy the eating couple for the new picture, which would have incurred an outrageous contradiction to the narrative content. The failed temptation of Adam seems to lead nowhere other than to the devil’s lure of Eve as an alternative tactic, which is, however, thematically distinct and lies too far ahead to be paired with the left half picture of the Temptation of Adam. Based on such reasoning, the artist would have followed the opposite direction of narrative sequencing and accordingly aligned the right half with a scene earlier than the Temptation of Adam, specifically the relatively long passage that precedes the serpent’s attempt at tempting Adam, namely, Genesis B 454a through 491b. In reversing the pictorial narration backward, he naturally would have taken into account the contextual connection to the immediately preceding picture, P13, to which P20 may well be reconceptualized as a thematic sequel. Through the interface of P13, furthermore, P10 and P11, particularly the former, would have come to the artist’s mind. (One may recall, in this connection, that these three pictures are the triad of the common theme, Adam and Eve in Paradise.) In order to make a smooth transition from P13 to P20, then, the Junius artist would have reproduced the happy state that Adam and Eve are situated in, as symbolized by the common poses of the couple in the two pictures. No less importantly, the artist would have represented Adam and Eve’s unified obedience to God, as described in the passage given above, drawing on the image of P10. In this way, the particular pose of the couple may be fully explained as echoing the preceding pictures in Junius in conceptual and visual terms. Raw (1953: 49) was apparently at a loss to account for the double figures at right in P20. All she managed to do was to suggest, without offering anything approaching explanation, that this duplication would have been a mistake for the single figuration as

170 

 7 The Temptation of Adam: P20/P24

realized at San Marco. Alternatively, Raw (1953: 49) referred to another possible error, which involves horizontal arrangement of scenes, rather than a figural representation as with the first mistake. Specifically, the upper right and left pictures in P20 may have been inverted by mistake in their sequence (for whatever reason). Then, the currently double-figured scene at right, which should have preceded the current left in the correct layout, would visualize an earlier event/state than the Temptation of Eve in harmony with the narrative development in the text. While Raw’s latter interpretation might seem similar to the one I have proposed above in terms of the sequentiality of events originally intended, the two differ markedly in their underlying conceptualizations. In my view, the artist never committed an error in arrangement—the extant right-to-left order would have been fully motivated in original—whereas he made such a grave mistake without obvious reason in Raw’s framework. In any event, according to Raw’s account, the artist erred incredibly and extravagantly either way, taking two figures for one on the one hand, and reversing right and left on the other. It should be recalled, furthermore, that the artist misidentified the left figure as Eve at some stage of manuscript transmission, whether one follows my account or Raw’s. Thus, according to Raw, gross errors were repeated one after another in this specific picture, whereas in my understanding a single mistake—if the alteration from Adam to Eve were so conceptualized—took place at a relatively late date far removed from the original artist’s composition. In conjunction with the Junius reconceptualization and reconfiguration of the original two pictures depicted in Grandval, the temporal sequencing of the emergent two scenes is determined as proceeding from right to left on inherent narrative grounds, which constitutes a reversal of the left-to-right unidirectionality that pervades in the majority of the Cotton Genesis family. Given that the reversal in horizontal orientation is implemented parallelly in other pictures in Junius 11—most notably in P9 (section 5.2)—in defiance of their inherited inverse directionality, it should be hardly surprising that a similar reversal is put into action in this case in the absence of structural reasons for avoiding the reverse order. For all credibility and explanatory power superseding earlier accounts like Raw’s or Broderick’s, the above proposal on the derivation of P20 is compromised in terms of evolutionary feasibility and relative complexity in derivation. Indeed, by conceptualizing the model as the antecedent of which the San Marco representation constitutes a transparent reflection, I carefully refrained from identifying the major source of the upper half of P20 with the San Marco Temptation of Eve itself. Yet it must be admitted that the preceding investigation is biased toward the San Marco testimony, although this bias is defensible in part. Aside from Millstatt, the San Marco mosaics provide the only example of the Temptation of Eve in its own right, thereby presumably witnessing, as in many other cases at San Marco, an otherwise lost earlier state of affairs. In light of the richest material that this artwork offers, it is no wonder that it strongly shapes comparative studies of the Cotton Genesis family, just as now. There is good reason, however, to believe that the composition reflected at San Marco in the present case deviates idio-

7.2 A diachronic–comparative perspective 

 171

syncratically from what would have been the Cotton Genesis archetype, thereby diminishing—if not disconfirming—the historical plausibility of the alleged model for P20. According to Weitzmann/Kessler (1986: 55, 131, pls. 6, 42), the tempted Eve appears at the center of the panel (at folio 9v as reconstructed, at folio 3v in the surviving Cotton Genesis manuscript), facing rightward. Correspondingly, the tempting snake—now damaged beyond recognition—in all likelihood would have been pictured at the far right. Such composition is closest to those actually embodied in Millstatt 10r (Pl. 63)— conflated with the Receiving of the Fruit—, the Hortus Deliciarum 17v (Pl.  69 upper left)—integrated with Adam Eating the Fruit—, and in the Salerno plaque (Pl. 56)—juxtaposed with the Fall of Adam and Eve. As far as the configuration of Eve and the snake is concerned, the Cotton Genesis scene of the Temptation of Eve constitutes a mirror image of what is represented in the Touronian Bibles, the San Paolo fresco (26r), as well as at San Marco. Furthermore, to the left of Eve is found additional space for distinct illustration. Weitzmann and Kessler (1986: 55) convincingly reason that the place would have been allotted to the episode of God Forbidding Adam and Eve to Eat the Fruit (God’s Admonition)—the only likely episode worthy of visualization—located in the left half of folio 9v between the Introduction of Eve to Adam (fol. 9r) and the Temptation of Eve (fol. 9v, right half), as realized in Grandval 5v2 (right, Pl. 24; section 6.1) and the Hortus Deliciarum 17r (Pl. 68). To schematically visualize the archetypal double-scene representation on fol. 9v in the Cotton Genesis by image-processing of the corresponding scenes in Grandval and the Hortus Deliciarum, one may arrive at the following two sets of reconstruction, in which the existence and location of the tree in the Admonition scene (left) may be regarded as immaterial, in the absence of agreement between Grandval and the Hortus Deliciarum in this respect. Common to both sets is the right-facing Eve in the right zone, as evidenced in the surviving fragment of the Cotton Genesis. Differing decisively is the location/orientation of the couple in the left zone: whether they appear just behind the right figure of Eve and are oriented left (Fig. 4), or they stand at the far left, facing right (Fig. 5). As far as the Admonition scene—the left half—is concerned, the first reconstructed image is independently attested in both Grandval and the Hortus Deliciarum, whereas the second one is unparalleled. Therefore, the first composition (Fig. 4) would appear more likely to have been embodied in the Cotton Genesis as well, although strictly in terms of probability the complementary distribution at issue lacks statistical significance warranting such probabilistic inference, obviously due to the small data size. Thus, in exploring the pictorial evolution leading to P20, one must start with the composition as reconstructed in Fig. 4 above, rather than the hypothetical antecedent of the Temptation of Eve at San Marco, as presupposed in the preceding account. In the earlier account, Eve’s Temptation of Adam/Adam and Eve Eating the Fruit in Grandval—the scene to the immediate right of the Temptation of Eve in Grandval 5v3—was singled out as a source for the right half of P20. This analysis requires reconsideration, considering that it has been established that the Temptation of Eve—to be adapted as

172 

 7 The Temptation of Adam: P20/P24

Figure 4. Approximation to the original image in the Cotton Genesis, fol. 9v, based on Grandval 5v (above) and the Hortus Deliciarum 17r and 17v (below). © The British Library Board: Add MS 10546 (Grandval Bible), fol. 5v (detail, adapted). Reproduction (detail, adapted): Green et al. (1979: fig. 21 and pl. 11). Credit: Warburg Institute.

7.2 A diachronic–comparative perspective 

 173

Figure 5. Alternative approximation to the original image in the Cotton Genesis, fol. 9v, based on Grandval 5v (above) and the Hortus Deliciarum 17r and 17v (below). © The British Library Board: Add MS 10546 (Grandval Bible), fol. 5v (detail, adapted). Reproduction (detail, adapted): Green et al. (1979: fig. 21 and pl. 11). Credit: Warburg Institute.

174 

 7 The Temptation of Adam: P20/P24

the Temptation of Adam in the upper left half of P20—was immediately preceded by the Admonition within the same picture panel in the Cotton Genesis. In other words, these two scenes were intimately connected in the original, whereas the Temptation of Eve is found immediately adjacent to Eve’s Temptation of Adam in the same frieze in Grandval as a matter of secondary development. For a diachronic–genealogical account of P20, two courses of action may be available in the present situation, one requiring minimal modification of the above proposal, the other deconstructing it wholly and formulating a radically distinct alternative. First, simply positing the original image reconstructed for the Cotton Genesis at a point in time earlier than the preform of the San Marco picture, one may exclusively address the putative additional change from the original Cotton Genesis image (Fig. 4) to what is attested at San Marco (Pl. 45), and leave the subsequent development to P20 unaltered exactly as presented in the earlier account. Second, starting from scratch, one may engage squarely anew with the evolutionary trajectory from the Cotton Genesis original to P20 and correspondingly review critically the alleged intermediary of pre-San Marco variant postulated above. Strategically speaking, the second option—a wholesale reexamination—seems by far the better choice, as it covers a broader range of problematics and at a deeper level of analysis than the first strategy (actually more of a makeshift than a genuine solution) can afford to address. While the San Marco mosaics, with their richest collection of images among the Cotton Genesis family (next to the Cotton Genesis itself in its original state), are usually regarded as more or less faithful reflections of the initial visualizations (cf. Weitzmann 1955: 119–121; Weitzmann/Kessler 1986: 18–20; Kessler 2009: 466), the assumption obviously needs qualification in the presence of direct testimony exceptionally retrieved from the surviving copy of the Cotton Genesis that is at variance with San Marco, much as with the idiosyncratic imagery in the Creation cycle peculiar to San Marco and unparalleled elsewhere in the Cotton Genesis tradition (cf. Kessler 2009: 471; 2009–2010: 21–23; 2014a: 11–14, 16; 2014b: 79, 81). The partially recovered picture of folio 9v in the Cotton Genesis is precisely a case in point. In methodological terms, therefore, precedence must be given to the Cotton Genesis (partial) image at the expense of the San Marco counterpart in identifying the original two-part visualization consisting of the Admonition and the Temptation of Eve. Needless to say, it remains still possible that this Cotton Genesis archetypal image underwent reversal along the horizontal axis at some point in transmission and subsequently served in this inversed form as a model for P20, among other descendants, on the one hand and San Marco on the other, as implied from the earlier claim. Accordingly, pursuing a novel account in accordance with the reconstructed Cotton Genesis image (Fig. 4), one must take this possibility—however remote—into consideration for a final assessment. As observed above, the Admonition (left) and the Temptation of Eve (right), standing side by side, appear in the same panel (fol. 9v) in the Cotton Genesis (Fig. 4). Thus, for whatever reason the two episodes constituted a pictorial pair in the beginning. Accordingly, this unique concatenation of images must be acknowledged as a point of depar-

7.2 A diachronic–comparative perspective 

 175

ture absolutely given, in exploring its subsequent evolution up to P20. In the majority of descendent cognate works the Admonition is rarely pictured, whereas the Temptation is ubiquitously depicted, usually combined with subsequent scenes, the Picking and Eating of the Fruit with Adam, that is, assembled and integrated as a whole into the Fall of Adam and Eve. In the course of losing autonomy in visualization, the Temptation scene more often than not was subject to reversal in horizontal orientation, whereby Eve stands at the right, facing left toward the snake, as in the Touronian Bibles, the Hildesheim door panel, the San Paolo fresco, as well as the San Marco mosaic. In light of such a wide-ranging distribution of the innovative mirror image crisscrossing geographical and genealogical subgroupings of the Cotton Genesis family, it may be plausible to assume that the image reversal occurred repeatedly and independently, rather than singly as a unique innovation. Given that the Grandval Bible exclusively retains the original juxtaposition of the Admonition and the Temptation of Eve (though separated in different registers), each standing on its own without being conflated with another episode, it may be warranted to give it privileged treatment, as it is minimally distanced from the initial state. In it, only the right half scene—the Temptation of Eve—was recomposed through reversal in face direction, whereas the left representation keeps its original orientation. This specific divergence in implementation of image reversal is subject to principled explanation. That is, all other logically possible combinations of operation/non-operation of reversal in the two halves—there are three other possibilities—are excluded as a matter of principle. As will be substantiated fully in section 12.3 below, God invariably appears to the left of man in Grandval (5v1–5v3), as does an angel (5v4). One may generalize that extraterrestrial beings are consistently aligned at the left in Grandval. Given such a generalization, any deviation from this general pattern of arrangement, as is the case with the Cotton Genesis Temptation image, would have been subject to regularization. Therefore, it is little wonder that the snake, Satan’s agent, be given an analogous place in a picture zone. In this way, the actually attested representation of the Temptation (Fig. 6) is derived straightforwardly from the original by the image reversal executed in conformity with the constraint on arrangement of extraterrestrial beings imposed on the Grandval Bible.106 In this respect, a comparison with the Salerno plaque (Pl. 56 right) would shed light on the issue. As pointed out above, the Salerno Temptation scene is one of the exceptional instances in which the tempted figure of Eve (left) is represented in her original rightward orientation. To its immediate right, the Fall of Adam and Eve is depicted in the same panel. Thus, Salerno and Grandval are minimally distinct: specifically, Salerno constitutes a mirror image of Grandval insofar as the left half—the Temptation of Eve— is concerned. In other words, Salerno was not affected by the reversal of orientation

106 For a somewhat similar but distinct motivation for the reversal of the serpent and Eve at San Marco, see Jolly (1997: 45).

176 

 7 The Temptation of Adam: P20/P24

Figure 6. Grandval 5v (detail), Admonition (left) and Temptation of Eve (right). © The British Library Board: Add MS 10546 (Grandval Bible), fol. 5v (detail, adapted).

that operated in Grandval. Such two-pronged evolution of the original Temptation may be readily explained by assuming that Salerno was not bound by the generalized constraint on arrangement of figures that governed Grandval with maximal consistency. The largely regular appearance of God at the far left at Salerno failed to apply with the maximal stricture characteristic of Grandval: angels may appear to the left of God (Pl. 53) as well as to the right (Pl. 52). In this light, it would be hardly surprising that the tendentious convention fell short of extending to Satan’s messenger and incurring change in face direction. With the Grandval scene derived from its Cotton Genesis counterpart through the strictly rule-governed mechanism elucidated above, I would propose that this instantiation of the Admonition and the Temptation in Grandval (Fig. 6) served as the sole model for P20, rather than a complementary pair of the pre-San Marco Temptation and the Grandval complex of the Temptation and the Fall, as formulated in the previous account. Evidently, there are a number of discrepancies between P20 and the mechanical juxtaposition of Grandval 5v2 (right) at the left and Grandval 5v3 (left) at the right (Fig. 6), apart from the altered identity of the person being tempted—textually motivated as expounded above—Adam (P20) versus Eve (Grandval). First, the order of the two episodes is reversed, right-to-left in Junius and left-to-right in Grandval (compare Figs. 6 and 7). This reversal has already been explained in a principled way above as part of the earlier account: as with other pictures in Junius such as P3.1 and P9, the original consistent left-to-right narrative progression in the Touronian Bibles was regularly transformed into the reverse order in Junius (for details, see section 12.3 below). The reordering is thus formally motivated on the principle of right-to-left linear progression distinctively characteristic of Junius. Second, while Adam and Eve face left in Grandval in conformity with the Cotton Genesis original as argued above, they are oriented in the opposite direction in Junius (compare Figs. 6 and 8). This repositioning/reorientation is no less motivated than the

7.2 A diachronic–comparative perspective 

 177

reversed narrative linearization. As will be discussed in the following sections and especially in section 12.3, Grandval and Junius are diametrically opposed with respect to the spatial arrangement of God. Specifically, God invariably appears at the left in Grandval as pointed out above, whereas he is placed at the right in Junius with comparable regularity. In accordance with this generalization, God would have been relocated to the far right in Junius (i.e., a Touronian Old Saxon antecedent of Junius 11), and correspondingly Adam and Eve were repositioned to the left, facing right toward God standing on the right. This prompts a third divergence between the two allegedly cognate scenes. While God admonishes the couple in Grandval, the corresponding scene in Junius does not contain a figure of God. Based on the couple’s orientation, however, it would be reasonable to assume that God would stand to their right if represented overtly. In Junius, God is occasionally represented in zero form; in other words, the figure of God is simply implied but not explicitly manifested, as in P31.2 (section 8.1) and P45.2 (section 11.1).107 By legitimating zero form as a variant of reduced representation in the pictorial system of the Old English/Old Saxon Genesis, and accordingly assuming a virtual presence of God beyond the tree across the page boundary, one may relate this picture to P10, God’s Blessing/Admonition (section 6.1), at a deeper level of pictorial organization, given that their intervisual correspondence has been substantiated by the remarkable resemblances between the two representations regarding the couple’s gesture (section 7.1). The above is the formal essence of the new proposal. The remaining minor adjustments in figural details such as the virtual superimposition of the couple in the right scene and Adam’s looking-back posture, as well as the textual alignment with the Old English/Old Saxon Genesis, require no change from the earlier account specifically presented above, as they must be invoked independently in any event. At this juncture, it remains to be argued how the new account supersedes the earlier one in substantive ways. First, the proposal claims greater historical plausibility in that it starts with the original picture demonstrably reconstructed for the Cotton Genesis and actually corroborated in Millstatt and the Hortus Deliciarum, without need to posit an unattested hypothetical antecedent of the San Marco image (Pl. 45) in the Cotton Genesis tradition. As detailed below, the San Marco image would have emerged as a consequence of an idiosyncratic recomposition additionally executed, as unparalleled elsewhere, regarding the presumable reduction of the couple to the sole figure of Adam. In this light, it is extremely unlikely that the unique compositional anomaly came to be shared in Junius, particularly in the absence of demonstrable contact between Junius and San Marco (cf. Broderick 2014: 223). Second, the new account is simpler in conceptual and derivational terms. All it requires is a single model: the succession of the Admonition and the Temptation in Grandval 5v. This single model thus dispenses with the complication of having to posit

107 On the zero representation of God as a variation of the reduced mode of depiction in the pictorial organization of the Old English/Old Saxon Genesis, see section 12.2.

178 

 7 The Temptation of Adam: P20/P24

a second source. It does not necessitate the extra derivational mechanism of merging the two figures, Adam and Eve, into one (in the case of the San Marco Temptation scene according to the previous account). Furthermore, since Adam and Eve face the same direction in the new model (the Admonition in Grandval), it is no longer necessary to invoke an extra operation of derivation and transform the confronting pair into the unidirectional one (as in the case of viewing the Fall in Grandval as a model). Third, the new alternative is capable of directly accounting for the synchronically motivated parallelism between P10 and P20.1 (right) by identifying the same model—the Grandval Admonition—as their common source. In other words, the earlier derivation cannot afford to construct a coherent account unified in the synchronic and diachronic dimensions. The lack of conceptual consistency and uniformity in the previous account is manifest in the reading of the right scene in P20. Since this picture allegedly originates from the Fall of Adam and Eve, it requires drastic reinterpretation in signification— from disobedience to obedience—along with the concomitant reorientation of Adam noted above. By contrast, according to the new proposal, the couple facing right immediately yields to a reading analogous to P10, in accordance with their common origin in Grandval, without the need to posit interpretive reversal. Finally, the unique representation of the Temptation at San Marco requires specific treatment. Its uniqueness concerns the exclusive appearance of Adam to the right of the tempted figure of Eve, turning his back on her. From a diachronic–comparative perspective, the couple is expected to appear instead (as in Grandval), or, if a single figure is chosen, it is Adam most likely to look toward Eve (as in Vivian). In this connection, one may be reminded that the Admonition fails to be depicted at San Marco as in many other cognates, with the Temptation immediately preceded by the Introduction of Eve to Adam (Pl. 44). In this light, it is reasonable to assume with Weitzmann/Kessler (1986: 55) that the single appearance of Adam constitutes a vestige of the Admonition scene.108 Such inference in turn may lead one to presume that the originally earlier episode of the Admonition was repositioned to the right of—as in Junius—and subsequent to—unlike Junius—the Temptation. This derivation, however, would defy principled explanation in terms of narrative linearization: San Marco obviously is not regulated by the principle of right-to-left narrative progression, in full force in Junius, that is held responsible for the reversal in question in the latter work. Concomitantly, the reversal did not occur in other cognate works (save San Marco) because the right-to-left sequentiality—the prerequisite of the reversal—is a privilege of the Old English/Old Saxon Genesis. By simply 108 Elaborating on Weitzmann’s (1984: 113) view that the Introduction of Adam to Paradise (Pl. 41) constitutes a conflation with the Admonition at San Marco, Jolly (1997: 26, 103n30, 44) attributes the dissociation of the Admonition from the Temptation in the same work, in defiance of their original juxtaposition in the Cotton Genesis, to its relocation to and conflation with the Introduction of Adam to Paradise. Whether or not such reordering actually occurred as Jolly contends, one must assume that, subsequent to the loss of the Admonition as an independent scene immediately preceding the Temptation, the original figure of Adam being admonished persisted, rather than totally eliminated, subject to reinterpretation in a new setting, as specifically argued below.

7.2 A diachronic–comparative perspective 

 179

Figure 7. Recomposition of the Cotton Genesis Admonition and Temptation, prior to Junius and San Marco. Reconstruction based on Grandval 5v. © The British Library Board: Add MS 10546 (Grandval Bible), fol. 5v (detail, adapted).

stipulating, in the absence of explanation, that for some reason or other an antecedent of San Marco, parallel to that of Junius, underwent the reversal in linear progression, one may obtain the composition given above (Fig. 7), using the Grandval material for reconstruction to facilitate overall comparison. The stage reconstructed in Fig. 7 may be characterized in purely formal derivational terms as commonly preceding Junius and San Marco. By the reversal of the right-side picture in accordance with the generalization that God should appear at the right, facing left, in Junius, one comes up with the representation in Fig. 8, which may be regarded as corresponding to a later preform common to both P20 and San Marco. With maximal reduction of the figure of God to zero, as motivated above for Junius, P20 would have emerged in the Touronian original of the current copy of Junius 11, concomitant with the essential reinterpretation of Eve as Adam and other accompanying adjustments. The San Marco image requires further operation before its eventual actualization, that is, the removal of Eve. This seems understandable, in light of the insight gained in the course of argument for deriving P20 from an earlier equivalent of the San Marco Temptation scene. That is, with rare exceptions (e.g., Eve Plucking the Fruit and Giving It to Adam, Pl. 46), the San Marco Genesis cycle strictly adheres to the principle of one scene per panel. Accordingly, in order to meet the compositional rule at issue, a full representation of the couple is avoided, as is a double depiction of Eve, thereby resulting unequivocally in a single figuration of Adam.109 In terms of derivational history, the San 109 With the original figures of God and Eve eliminated, the remaining Adam may read as turning away from Eve being tempted by the snake. God’s Admonition no longer appreciable, the novel scene is subject to reinterpretation along the line that Jolly (1997: 47) contends to be the case: “. . . but here the mosaics assert that Adam is ignorant of any such encounter or, if he hears it, turns away. The mosaicists, by including an uninterested Adam rather than simply omitting him, powerfully reinforce his uninvolvement.”

180 

 7 The Temptation of Adam: P20/P24

Figure 8. Recomposition of the Cotton Genesis Admonition and Temptation, immediately preceding Junius. Reconstruction based on Grandval 5v. © The British Library Board: Add MS 10546 (Grandval Bible), fol. 5v (detail, adapted).

Marco image is thus farthest removed from the original, as it is differentiated from the rest of the Cotton Genesis family by the sole presence of Adam looking away from the tempted Eve.110 Given the above insight into the derivational processes involved, the earlier account proves to have erroneously identified an evolutionary later instantiation as a direct model for a genealogically earlier cognate. Put another way, the previous proposal of deriving P20 from an antecedent of San Marco contradicts their presumable relative chronology, whereby the older item was allegedly shaped by the newer one. This fallacy inherent in the otherwise apparently plausible account has been brought to light by exploring Weitzmann and Kessler’s reconstruction of the Cotton Genesis folio 9v to its logical conclusion. Before moving on to P24, it may be worth mentioning that, as suggested above, the tree in the center in P20, in addition to serving as the demarcation between the two halves, plays the double roles of representing the Tree of Death for the right scene and the Tree of Life for the left. Thus, Adam and Eve are turning away from it in obedience to God in the right scene, whereas Adam is resisting the devil’s temptation by restraining his left hand from taking fruit from the left tree and turning instead to the Tree of Life at the center. In addressing the sources of the Temptation of Adam by Satan’s messenger in an angelic form depicted in P24, one may well start with a critical appraisal of Raw’s (1953: 50) claim. One should be reminded at the outset that Raw, like many others, takes at face

110 Based on this formal derivational relationship, one might be prompted to speculate that the San Marco image was actually derived from the Junius composition. For further discussion on the presumable influence of Junius on San Marco, see section 12.4 below.

7.2 A diachronic–comparative perspective 

 181

value the identity of the figure being tempted as Eve: the artist responsible for the figuration would have made such identification in conformity with the biblical convention and conversely in violation of what is narrated in the Old English/Old Saxon Genesis text. Accepting this resultant disparity between picture and text as an absolute given in the very beginning, one may find no further interest in exploring implications that this pictorial mismatch entails for a deeper understanding of the synchronic and diachronic dimensions of manuscript organization and evolution. With an initial observation that the Temptation of Adam or Eve by an angelic being is a unique feature of Genesis B, Raw proposed to treat the angel in P24 (and also P28 and P31) as a genealogically extraneous, hence, equivalent to nonexisting or invisible being, and concentrate accordingly on the composition of the remaining figures for comparative inquiry. As a consequence of such a heuristic operation of deletion, Raw arrived at the finding that P24 constitutes a mirror image of Grandval 5v3 (left) and San Marco (Pl. 45), for which the serpent—absent as it is in P24—is correspondingly bracketed for purposes of comparison. Thus eliminating the angel from P24 on the one hand and the serpent from Grandval 5v3/San Marco on the other will lead to the identification of the pictures at issue as isomorphic, and therefore equivalent to each other, with respect to the configuration of Eve and the facing tree, and in disregard of the orientation of Eve, facing leftward (Grandval and San Marco) or rightward (Junius). Accordingly, P24 is derived from the traditional scene of the Temptation of Eve by the Serpent—as represented in Grandval 5v3 and San Marco—through a series of figural transformations. The operations invoked include a horizontal inversion, a deletion from the tree of the original tempter (serpent), and an insertion, between Eve and the tree, of another figure of a tempter (angel). While Raw’s account admirably revealed the pictorial isomorphism and the formal mechanism of derivation, it remains unaddressed what cognitive forces would have been at work to drive the artist to put such formal operations into practice for creating new scenes. Indeed, it may be quite intelligible that in structural and mechanical terms a mirror copy may provide a viable and economic means of image multiplication such that it may obviate a separate search for an independent model elsewhere. Yet, in addition to such a formal advantage, more specific, substantive motivations must have been in force to convince the artist to decide on the actual implementation. At this point, one must bring the hypothesis submitted earlier on the emergence of the text–picture incongruence regarding the Temptation of Adam/Eve in P20 and P24, and particularly the account presented above of the upper part of P20, to bear on the cognitive issues raised. As may be recalled, the underlying claims proposed here are that the pictorial mismatch arose only secondarily relatively late in the course of transmission, and accordingly that the figure represented as being tempted in P20 and P24 would originally have been designed to be Adam in perfect harmony with the text of the Old Saxon antecedent of Genesis B. As argued above, had it not been for a complete match with the vernacular text in regard to the Temptation of Adam, the right-facing (i.e., turning away from “Eve”),

182 

 7 The Temptation of Adam: P20/P24

largely overlapped, double figures at right in P20 would be hardly susceptible to principled explanation. In other words, it is reasoned, the two overlapping figures on the right would imply that the artist must have been sufficiently familiar with the narrative content unique to Genesis B to rework and reorganize the right scene as a visualization of Genesis B 454a through 490b. From this, it may be but a short step to inferring with reasonable certitude that he was well versed in the following long passage in the text as well, beginning with Genesis B 491a (maximally through 546b). Thus, with full knowledge of the Old Saxon text underlying Genesis B, the original illustrator would have reworked, in perfect concord with the vernacular narrative, the image sequence reversed from the original succession of the Admonition and the Temptation in the Grandval Bible. In light of such a radical reconfiguration of the model pictures, it may be scarcely surprising that the artist also would have implemented another analogous reconfiguration—a horizontal reversal relative to P20 (upper left)—to obtain a framework for P24. One may still wonder why the artist decided to reverse the orientation at all, although it is now clear that he would have been at liberty to alter directionality. Here it may be illuminating to refer back to the observation made earlier in the preceding section that P20 and P24 are differentiated primarily by changes in perspective, viewed from this side or from the opposite side of the scene beyond. Complementary to this spatial change in the viewer’s orientation is a slight shift on the temporal axis. Specifically, one may imagine, in a brief transition period from the left side of P20 to P24, the serpent leaves the tree—the Tree of Death—and stands in front of it, whereupon it takes on the figure of an angel. Thus, the two scenes, P20 (upper left) and P24, may be viewed as two phases of the same episode that are slightly shifted one after the other on the temporal axis. It is during the relatively long confrontation scene after this moment of metamorphosis that P24 fixes a point for visualization, but seen from the opposite perspective to the one taken for P20. Reconstructing the composition of a new image in the artist’s mind in this way would provide a cognitive basis for the horizontal reorientation that is assumed to have been executed in P24, in contrast to P20. It should be noted that this understanding is made possible by an integrated conceptualization of P20 and P24, which builds on the assumption that both pictures were originally conceived as an inseparable unit with the Temptation of Adam in full accordance with the text of the Old Saxon original of Genesis B. While P24 is conceptually derived from P20 (upper left) as reconstructed in cognitive terms in the preceding paragraphs, it does not mean in the least that this formative process would have proceeded exclusively in the mind without associating with the perceivable world outside. Put another way, the image of P24 is by no means a product of purely mental power working in the abstract. There is another dimension to the image construction, based as it is on concrete material. Specifically, in shaping a new image, the artist would have drawn on preexisting sources for concretization and elaboration of his mental image. In other words, as in other cases he presumably would have consulted existing examples as his models. In view of the fact that the tempter in the form

7.2 A diachronic–comparative perspective 

 183

of an angel is nowhere to be found in the biblical tradition, however, one should better refer to them as first approximations of the artist’s mental images. At any rate, regardless of the naming, the artist was open to inspiration from whatever sources he considered most relevant to his concerns. Given that the original artist intended the human figure to be Adam according to the account proposed above, and also given that the configuration of Adam (or Eve for that matter) and an angel in the temptation scene is unprecedented in the tradition of biblical illumination (cf. Raw 1953: 49), he would likely have turned for consultation to the extant images in which Adam is featured centrally and accompanied with some other being as well. Furthermore, since the figure assumed by Satan’s messenger is an angelic being, its representation more likely would have been inspired by other celestial beings independently accessible in the illustrations, figures that the artist was led to conceptualize as comparable to his conceived image on analogical pictorial grounds. By contrast, in the absence of the tempter as an angelic being to model after, the artist hardly would have been convinced of adopting the traditional scene of Eve tempted by the serpent as exclusively the most useful concrete material for his creative reuse. Thus, as far as the existing materializations are concerned (apart from what is going on in the mind), the Temptation of Eve by the Serpent unlikely would have been counted as self-evidently optimal due to the discrepancies in the identities of both the tempted (Adam vs. Eve) and the tempter (angel vs. serpent). It is only later at the stage of revision and copying—in which the original figure of Adam was modified as Eve—that these scenes come to assume paramount importance as determining forces for the reconstitution of Eve as the tempted figure in replacement of Adam. Following the above lines of reasoning, then, it may be inferred that artistic inspiration is likely to have been provided by the Animation of Adam, in which Adam, having been shaped materially, is given spirit by God. While the scene recurs in several cognate works (Kessler 1977: 15–16), what seems to be corresponding most closely to P24 may be recognized in San Paolo 8v1 (the middle panel in the top register; Pl. 34) and/or San Marco (Pl. 40), in both of which Adam at left and God at right stand face to face, precisely as in P24. In San Paolo, wearing long hair and touching his left hand with his right (which is not raised, though, in distinction from P24), Adam is shown standing in front of God, who raises his right hand toward him. Adam and God are situated between the trees in the San Paolo scene, as are Adam and Satan’s messenger in P24. Further, a short thin plant grows between the two figures, again in parallel to P24. In the absence of these surrounding features, the San Marco scene, which exclusively focuses on the two standing figures of Adam (at left) and God (at right) for representation, appears less similar at first sight. An outstanding feature unique to this work, however, is the explicit depiction of a spirit in the form of a tiny winged figure, handed from God to Adam, the configuration analogous to the treatment of the fruit in P24. These complementary agreements in fine pictorial details may therefore warrant concluding that P24 would have been inspired in no small measure by the Animation of Adam scene exemplified in San Paolo 8v and at San Marco.

8 The Fall of Adam and Eve: P28/P31 8.1 A synchronic–structural perspective Satan’s messenger disguised as an angel in P28 (Pl. 14) may be reasonably regarded as a common agent responsible for offering the fruit to Adam and Eve, who react divergently to the temptation: while Adam refuses to take it, Eve accepts and eats it. Indeed, this overall interpretation hardly seems contestable (cf. Gollancz 1927: xlii; Broderick 1978: 208; Ohlgren 1992: 91), but it is subject to two distinct readings differentiated by their temporal specifications. First, as may appear most natural at first glance, the devil gives the fruit to Adam and Eve at the same time. This seems unlikely, however, given what is narrated in the text. Yet one may refine a nonsequential interpretation by assuming that P28 represents in summary fashion and thus in atemporal static terms the agent’s successful and failed temptation of Eve and Adam, respectively. As illustrated in the following picture (P31, Pl.  15), it is Eve who successfully persuades Adam to eat the apple. Alternatively, differentiated along the parameter of temporal sequentiality, the messenger offers the fruit to Adam first and then Eve on a separate basis, as proposed by Ohlgren (1972a: 207).111 In analogous fashion, one may further view Eve’s representation as an amalgamation of two acts: she receives the fruit from the messenger in her left hand and then eats it with her right, rather than holding two pieces in both hands. In a way, such a merge of two distinct scenes involving the couple Adam and Eve is comparable to P44 in which God is depicted twice, however, rather than being integrated into a single figure (section 10.1). Thus, P28 constitutes a complete amalgamation, whereas P44 is characterized as a case of juxtaposition. For a fuller analog of P28, see below. The logically possible reverse order, Eve offered first and Adam second, should be ruled out as it contradicts the narrative development in the text. Thus, when taken in isolation, both the temporal and atemporal interpretations seem to make sense insofar as one does not adhere to simultaneity or to the left-to-right successivity in actions. The assessment of the two competing accounts must await examinations explored in a broader context, as substantiated in the following paragraphs. Bringing in other Junius pictures for comparison will shed a new light on formal and conceptual properties of P28, a perspective that has hardly been taken systematically in previous scholarship. As it turns out, P28 is implicated in outstanding parallel-

111 “The Tempter holds an apple in each hand, appearing to offer them to Adam and Eve at the same time. If we focus on the right half of the drawing, it seems apparent that the Tempter first offers the fruit to Adam, who rejects it. As the focus of attention shifts to the left half, we note that Eve accepts the apple with one hand and eats it with the other. The repetition of the apple indicates a sequential drawing, whereby several moments of the story are depicted as one scene without repeating any of its participants” (Ohlgren 1972a: 207). See also Karkov (2001: 70). https://doi.org/10.1515/9783110788068-008

186 

 8 The Fall of Adam and Eve: P28/P31

ism to and marked contrast with P13 (Pl. 9). Set within a similar architectural frame, Adam and Eve are confronted, directly (P13) and through the intermediary of Satan’s messenger (P28). Adam and Eve’s positions relative to each other are reversed, Adam standing on the right in P28 but located on the left in P13. While many animals and trees are depicted in P13, there are none whatsoever in P28. Instead of the prominent figure of Satan’s agent at the center in P28, a third figure enclosed in a mandorla in P13—identified as God according to Gollancz (1927: xli) and Ohlgren (1992: 90), particularly on account of the accompanying rays of light (section 6.1)—is represented in a high position and a small size between the arches. This profile bust of God in a mandorla issuing rays of light in P13 constitutes a stark opposition to the black background of the acanthus decoration set in the corresponding place between the arches in P28. Noteworthy in this connection are the white background against which the similar figures appear above Adam on the one hand and the corresponding blank space above Eve on the other in the same picture. Deserving further attention are the black thornlike projections pointed downward from the frame overhead, which may be viewed as pointing to the demonic nature of the devil agent assuming an angelic form for deception (Broderick 1982: 40). Not to be dismissed either is a group of geometric figures in the upper frame in P13 in opposition to the vegetative figures in the corresponding place in P28. The geometric figures in question are similar to those on the upper frame of the door above the angel in P46 (Pl. 22) and less outstandingly to those above Adam and Eve leaving paradise in P45 (Pl. 21). The vegetative figures on the black background accordingly may be construed as a symbol of the gate to hell, the dark world, in contrast to the geometric (or celestial) ones symbolic of the entrance to heaven. While P28 is thus analogous in many respects to P13 in compositional terms, the two pictures differ fundamentally in regard to the inherent dynamics in their pictorial narration. P13 represents a single scene, whereas P28 constitutes a complex of two distinct events by conflation, as pointed out in Ohlgren (1972a: 207) and elaborated above. More specifically, apparently integral as a whole, P28 falls into two parts, the right and left halves. Each half is a representation of Satan’s agent tempting Adam (right) and Eve (left), respectively, as interpreted above. The compositionality of P28, and the meanings of its separate parts, will be revealed more clearly when one invokes other pictures in Junius 11 for a fuller accounting of the apparently simplex illustration. This eventually will choose the right-to-left temporal reading at the expense of the atemporal summary representation discussed earlier. The right side of P28 corresponds to P24 and less obviously to P20 (upper left): Adam refuses to receive the fruit from Satan’s agent. Correspondingly, the left half of P28 is in contrast to P24 and to the upper left half of P20: while Adam refuses to eat the fruit (his right hand holding the elbow of his left in P20 and P24), Eve eats one with her right hand while receiving another fruit in her left hand, or more plausibly Eve eats the fruit with her right hand after having received it in her left, as argued above. Eve’s posture is comparable to Adam’s pose in P24, with the decisive difference being the action of the right hand, eating versus refusing the fruit. This contrast becomes most pronounced

8.1 A synchronic–structural perspective 

 187

when one views the left half of P28 in turn as comprising two successive acts, receiving and eating the fruit. The two constituent scenes are sequenced from right to left, as Adam is approached by Satan’s messenger earlier than Eve according to the Genesis B text. Thus, the right-toleft order is implemented in conformity with the conceptually related pictures that are also temporally differentiated on the horizontal axis, namely, P3.1 and P20 (upper).112 Compositionally, the messenger (P28) is highly reminiscent of Lucifer (P3.1): both look back with the arms spread out wide. A common message underlying such a compositional parallelism may be detected to the effect that “at left is represented the exemplar, which the figure(s) at right are urged to associate with.” Of greater significance, both pivotal figures may be characterized as conflations of the figures appearing in separate scenes. Analogous to Satan’s messenger, the central figure of Lucifer in P3.1 belongs to two distinct scenes, as proposed in section 3.1 above: on the one hand Lucifer is inviting his followers to the new palace, and on the other he is crowned and enthroned in the new palace and being approached by a procession of his subjects offering him crowns. Alternatively, as suggested by Broderick (1978: 208), the whole scene of P28 might possibly be read as a series of successive events in which Eve, having received the fruit from Satan’s messenger and eaten of it, hands the fruit over to Adam through the intermediary of the devil agent, which he refuses to accept. However, this sequential interpretation, for all its intricacy, does not seem to be particularly illuminating in moving the viewer’s eyes. Proceeding from center to left, and then to extreme right past the center necessitates a complex bidirectional shifting of the viewer’s focus that is unprecedented in other pictures in Junius 11. Moreover, this narrative reading is contradicted by the upper picture in P31, in which it is Eve, not the messenger, that physically gives the fruit to Adam, in accordance with the Genesis B text (717b–719b). In this light, the right half of P28 must be identified as the scene of Adam refusing to receive the fruit from the devil agent well before Eve is deceived, as represented by way of confirmation and reinforcement in other variant pictures (P20, upper left, and P24). P46, the Expulsion of Adam and Eve (Pl. 22), is another picture that is associated with P28 through a number of similarities and contrasts (section 11.1). Instead of Satan’s messenger, an authentic angel appears in P46, not at the center as in P28, but on the left periphery. The location of the devil and the angel naturally correspond to the close pairing of Adam and Eve (P46) and the splitting of the coupled configuration (P28). As in P28, the angelic being is represented most prominently among the three figures depicted in P46. Furthermore, these two extraterrestrial beings assume a highly similar posture, with arms spread wide apart. While the devil stands on the same level as Eve (but lower than Adam), the angel occupies a higher position than Adam and Eve, as does God

112 This formal parallelism may be adduced as a further argument against the alternative reading involving the complicated trajectory of the fruit moving first from center to left and then to right, as criticized in the following paragraph in the text.

188 

 8 The Fall of Adam and Eve: P28/P31

usually. In both, Adam and the extraterrestrial being face each other, to the exclusion of Eve, who is oriented in the opposite direction to Adam, either confronting him (P28) or glancing away from him (P46). The architectural frame, commonly existing above angelic beings, differs markedly in decoration: the acanthus on the black background (P28) is contrasted sharply with the geometric figures (P46), in parallel to the opposition between P28 and P13 discussed above. Divided into two parts sequencing vertically from upper to lower, P31 (Pl. 15) represents two successive scenes that take place immediately before and after Adam’s eating of the fruit (Gollancz 1927: xlii; Ohlgren 1992: 92; Broderick 1978: 211–212), the critical act that is not illustrated in itself. Above (henceforward referred to as P31.1), Eve humbly offers the fruit in both hands to Adam, who receives it in his right hand.113 Behind Eve stands Satan’s agent, evidently spurring Eve to tempt Adam. Below (P31.2), both kneeling on the ground, Adam looks up and raises both hands in prayer, and Eve faces down to the ground with her hands about to cover her face in despair. Behind the couple stands Satan’s messenger, now turned back into a demon’s original form and triumphant after having accomplished his mission. The upper and lower halves outlined above contrast sharply in their manner of configuration, although the same three beings are depicted in the scene. Specifically, as observed by Broderick (1978: 215–216), the messenger appears at the left in the top picture, but at the right in the bottom one; in similar vein, Adam is seen to Eve’s right and left in the upper and the lower zones, respectively. To generalize Broderick’s pairbased observation, the linear placement of the three figures involved is reordered in the two panels of P31, whereby the periphery members—the devil and Adam—were switched in their position but Eve remained constant in the middle. Alternatively, one may conceptualize that the whole concatenation of the three figures is subject to reversal: the sequence of Satan’s agent–Eve–Adam (left to right) in P31.1 is converted to that of Adam–Eve–Satan’s agent in P31.2. While Broderick’s description reductively characterizes the resultant reversal as a consequence of a local pair-based change, that is, the swapping of two ends, the alternative view offers a holistic account of the overall change as involving the entirety of the three figures. This difference in perspective should be assessed in cognitive terms rather than viewed purely as a formal variation of a conceptually equivalent (or neutral) generalization. Specifically at issue is which of the two competing accounts can make better sense of the artist’s presumable motivation for implementing the change in linear order in P31.1 and P31.2. By formulating the reversal holistically as affecting all three constituents, one may be able to conjecture that the wholesale transformation in the entire order would have been designed as an iconic visualization of the upheaval that befell the fate of mankind, as specifically represented in P31 and the following pictures.

113 Raw’s (1953: 49) description, as follows, must be rejected simply as erroneous: “The angel stands at the left of the picture behind Adam, who is offered an apple by Eve; he takes it in both hands.”

8.1 A synchronic–structural perspective 

 189

Figure 9. Compositional parallelism between P24 (left, horizontally reversed) and P31.1 (right). Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 24 (detail, adapted), Temptation of Adam; and p. 31 (detail), Eve Tempts Adam. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).

As elsewhere, systematic comparisons with other pictures in Junius may contribute to unraveling deeper meanings of P31, by situating it in an interpretive network of pictorial images from synchronic perspective, than a first approximation based solely on an individual and isolated inspection. The figure of Adam in the upper part of P31 is almost identical to that in P28, the sole difference being that while Adam does not touch the fruit in P28, thus rejecting it, he receives it from Eve in P31.1. Correspondingly, Adam finally succumbs to temptation in P31.1, whereas he withstands it for the time being in P28. It also should be noted that Adam’s posture in P31.1 is no less comparable to that in P24. A horizontal reversing would highlight the figural similarity with maximum effect (Fig. 9). Accordingly, one might be misled into viewing the figure in P24, too, as having fallen prey to the deception, and hence conclude that the deceived figure must be Eve in accordance with the biblical tradition—the argument that might possibly call into question anew the claim presented above that the figure must have been originally intended to be Adam. To be sure, the two figures in question touch the fruit, but one should face two significant differences squarely. First, while Adam receives the fruit in his right hand in P31.1, he touches it with his left in P24 (Pl. 13). Second, the right hand holds the left elbow in P24, apparently a restraining gesture as discussed in section 7.1 above, whereas the corresponding left hand does not restrain the right one in P31.1, pointing instead toward Eve or the fruit. Particularly in light of the latter contrast, it seems legitimate to make a crucial distinction between the two apparently similar figures: one may then conclude, while Adam is shown resisting the devil’s temptation in P24, he succumbs to Eve’s persistent enticement in P31.1. In addition to the similarity in Adam’s pose, the emissary’s posture in the upper picture of P31 closely resembles that in P24. While mirror-imaged, the demonstrable overall resemblance has a cognitive basis: both figures commonly urge the acceptance of the fruit by deception. Of greatest significance among the oppositions between P31.1

190 

 8 The Fall of Adam and Eve: P28/P31

Figure 10. Lucifer in P3.1 (middle) and his messenger in P31.2 (left, horizontally reversed) and P28 (right). Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 3 (detail), Fall of the Rebel Angels; p. 28 (detail), Temptation of Eve; and p. 31 (detail, adapted), Adam and Eve Pray in Remorse. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).

and P24 is evidently the presence versus absence of Eve between Adam and the emissary. Reversing either picture along the horizontal axis leads to an almost perfect match, except for the intervening figure of Eve (Fig. 9). The absence of another tree behind the devil agent in P31.1 may be readily attributed to the presence of Eve, which eliminates the marginal space otherwise available for drawing the tree behind the devil. The identification of P31.1 and P24 in terms of structural configuration thus postulated may in turn reinforce the original identity of the human being in P24 as Adam rather than Eve. The conventional account of P24 as the Temptation of Eve would fail to appreciate the pictorial–cognitive equivalence otherwise in evidence.114 Moving on to P31.2, one may immediately be struck by a postural isomorphism to Lucifer (P3.1) and, as mentioned above, to Satan’s messenger disguised as an angel (P28). Once again, a horizontal reversal that may operate on P31.2 by simulation would reveal to the full the virtual identification involved, as illustrated in Fig. 10. The mirror-imaging implemented on P31.2 as against P3.1 and P28 may be subject to an explanatory account: as argued above, what stands to the left of God’s antagonist and is pointed to by the fiend is presented as something that he has set as his goal to strive for: the throne in a new lofty palace (P3.1) and Eve’s yielding to temptation (P28). Seen in this light, Adam and Eve’s fall, the ultimate goal of the actions the emissary has been devoted to, should be represented at the left side of the page. Since the two humans must be aligned with the

114 Notice that Adam’s hairstyle is closely matched in the two pictures at issue, despite the fact that he is often depicted with short hair, as in P20 (upper right), P28, and P31.2.

8.2 A diachronic–comparative perspective 

 191

left, the remaining figure, Satan’s agent, is placed on the right periphery, as is actually materialized, rather than at the center of the page. The kneeling Adam and Eve in P31.2 may evoke a similar image in the lower part of P41 (P41.2, Pl. 19). In P31.2, Adam and Eve are praying to God in heaven, who is absent from the scene in paradise yet whose presence is implied far away to the upper left. In place of God, his opponent is shown at the right. By contrast, in P41.2, with the demonic emissary gone back to hell, God is now situated in paradise and calling to Adam and Eve, who are hiding among the trees (section 10.1). The figures of Adam and Eve are further paralleled down to the details of their pose: Eve is kneeling more deeply than Adam in both pictures. Also noteworthy is the tree that stands between Adam and Eve in both illustrations. These two analogous pictures, however, exhibit two significant differences concurrently. First, as described immediately above, while God is present on the spot in P41.2, he is absent away from paradise in P31.2, in which the presence of God high in heaven is suggested by Adam’s face orientation. Second at stake is the differing initiator of communication. With Adam and Eve addressing God in P31.2 on the one hand and God speaking to Adam and Eve in P41.2 on the other, the agentive role of a communicative act is switched in the two pictures. Building on these complementary properties, one may venture to assume that P31.2 is causally linked with P41.2. That is, God, addressed by Adam and Eve in prayer (P31.2), calls back to them in reply (P41.2), apparently after a lapse of time in the course of events as represented in P34, P36, and P39. Compare Genesis A 852a–854b: “þā cōm fēran / frēa ælmihtig // ofer midne dæg, / mǣre þēoden, // on neorxnawang / nēode sīne” (Then after midday the Lord almighty, the famous prince, came to walk in paradise at his leisure; Anlezark’s [2011: 65] translation). In other words, P31.2 and P41.2 are paired as a set of call and response between man and God.

8.2 A diachronic–comparative perspective In line with the foregoing (and forthcoming) explorations into other pictures in the present book, one may conceptualize the overarching organization and configuration at a higher or more abstract, significant level of visualization than the one concerned with more concrete features of individual figures. At this level of representation, it should not necessarily hold that individual identities are faithfully inherited from a model figure to its figural reflex in a derivative work. Of overriding importance at this level of abstraction is a formal–structural equivalence and parallelism in pictorial organization. From the more or less conventional perspective of prioritizing individual figures and their attributes, what is primarily at issue is the correspondence of features or lack thereof in the figures being compared, under the condition that the entities involved constitute the same identities, Adam and Eve in the case at hand. Since Satan’s messenger transfigured as an angel is a novel character unknown to the biblical tradition, one may legitimately leave it out of consideration for the sake of comparative inquiry, the

192 

 8 The Fall of Adam and Eve: P28/P31

guiding principle that Raw (1953: 50), for example, most explicitly articulates and vigorously applies. From such an individualistic viewpoint, it is exclusively Adam and Eve in P28 and their configuring that are selected for examination, with the central figure of Satan’s agent being treated as extraneous to the meaningful—in terms of comparative inquiry—image space as a whole. Accordingly, P28 is most convincingly compared to the pair of Adam and Eve in the central scene of Grandval 5v3 (Pl. 24) particularly by virtue of the close-resembling pose of Eve eating the fruit, and, slightly less conspicuously, to the corresponding pair at San Marco (Pl. 46), in which Eve is not shown eating the fruit. Such is essentially a view that seems commensurate with Raw (1953: 50) and Broderick (1978: 209), and that comes closest to being a consensus, although there may be a range of subtle differences in points of focus in theorizing and paths of derivation in reconstruction. Complementary to the above perspective of individuation, one may take into account a higher level of composition by looking into larger entities of configuration and their organization without being constrained by invariance in the identities of constituent figures. Attaching uttermost importance to Eve’s posture of raising both hands that seems strikingly similar to the pose of Satan’s agent in P28, one may assume that the three concatenated figures of the snake, Eve, and Adam in Vivian 10v2 (Pl. 27) and Bamberg 7v2 (Pl. 23) would have been recomposed through adaptation as those of Eve, Satan’s agent, and Adam in a novel configuration of P28. Or should one place the greatest emphasis on the dual directionality—extending the hands both forward and backward—and hence the two-facedness of the central figure, one may alternatively claim that the configuration of the snake, Eve (twice depicted), and Adam in Grandval 5v3 (left, Pl. 24) and San Marco (Pl. 46) would have been reorganized as that of Eve, the emissary (through merging the two figures of Eve), and Adam, whereby Eve and Adam facing each other in Grandval and San Marco are reflected in the devil’s glancing back to Adam, who is looking forward in P28. In corresponding fashion, the left-sided figure of Eve touching the fruit in Grandval and San Marco is reflected in the devil’s right hand holding the fruit in P28. A notable advantage of the derivation (in either version) of the three figures in P28 from the sequence of figures ranging from the snake to Adam in Vivian, Bamberg, Grandval, and San Marco concerns the vegetative decoration most conspicuously displayed precisely above Satan’s messenger in P28. This seemingly coincidental feature may be ascribed to the exuberantly grown tree that the snake entwines in the model pictures. By virtue of the conceptually inseparable link between the tree and the snake—the earlier guise of the emissary—the vegetative attribute may have accompanied Satan’s agent in the latter’s relocation—with transfiguration—to the center. Since a tree constitutes no comparably conspicuous presence in composition in the alleged models—the single scene of Eve giving the fruit to Adam in Grandval 5v3/San Marco— postulated in the earlier accounts (e.g., Raw 1953 and Broderick 1978, as mentioned above), the pictorial correlation of the devil to the vegetative decoration in P28 must be left unaccounted for.

8.2 A diachronic–comparative perspective 

 193

Whichever particular version one may eventually adopt, the identification of P28 with the complex of scenes—the Temptation of Eve by the Snake and the Temptation of Adam by Eve—in the Touronian Bibles and the San Marco mosaics as proposed above commands a greater explanatory power than the conventional view of associating it solely with the Temptation of Adam by Eve. Since the Genesis B scene of Adam refusing and Eve yielding to the devil’s temptation (Genesis B 495a–599b) does not necessarily require the arrangement of the three figures involved exactly in the actual way that is strikingly similar to the compositions that were executed in the Touronian Bibles for other related but distinct episodes, one must infer that the overall resemblance in configuration is far from a pure contingency; the most plausible account accordingly would be to assume that the close compositional similarity at issue would have been motivated on genealogical grounds. With the overall justification and rationale thus articulated, one may well enter into specific details in the proposed derivation. To start with the central figure in P28, it must be explained why the devil ends up turning back to Adam rather than looking forward to Eve.115 This is a question that deserves serious consideration for both the competing accounts. On the one hand, in Vivian 10v2 and Bamberg 7v2, the central figure of Eve looks toward the snake at the left. Accordingly, the reversal of face orientation needs to be explained in the course of derivation. On the other hand, inasmuch as the two central figures of Eve in Grandval 5v3 and San Marco look in opposite directions, the eventual selection of one over the other in P28 still demands explanation. At this point, one may invoke the pivotal figure of Lucifer in P3.1 as an explanatory model for P28 and less conspicuously for P31.2 as well (see below). As argued in the previous section, the figural resemblance of the central figures in P3.1 and P28 (Fig. 10, section 8.1) is conceptually motivated: both Lucifer and his emissary urge the angelic or human beings standing on the right to join what is presented to be the exemplary state actualized on the left side. Reinforced further by the metonymic/metaphoric relation involving Lucifer and Satan’s agent, this speech act of enticement commonly executed by both figures would likely have induced the artist to orient the devil backward to Adam at the right in P28 in line with Lucifer’s analogous pose in P3.1. In other words, the artist would have been inspired and influenced by his own creations as well as inherited models. Assuming that any instantiation presupposes formal and cognitive underpinnings for materialization, one may make a further generalization and abstraction by conjecturing that the artist did not mechanically and literarily follow P3.1; rather, he would have been guided by the underlying pictorial–cognitive schema for representing a tempter at large, of which P3.1 is a representative manifestation. As far as the left figure of Eve in P28 is concerned, apparently one must postulate a replacement of the snake by Eve. In this respect, adherents to the traditional account

115 Compare Ohlgren’s (1992: 91) observation: “The devil’s head is turned toward Adam, which may indicate that he is being tempted first . . .”

194 

 8 The Fall of Adam and Eve: P28/P31

may well argue, their conventional account needs no special pleading for such a figural replacement, since the identity of Eve remains unchanged through inheritance from the Grandval and San Marco models represented as the Temptation of Adam by Eve. This alleged gain, however, must be paid by giving an extra account of the insertion of Satan’s emissary between the otherwise contiguous figures of Adam and Eve. On closer examination, however, this substitution of the snake by Eve does not need to be posited as an independent process of adaptation; rather, it is a logical consequence of the redefinition of the central figure as the emissary. Because Adam’s position remains fixed at the right side of the page and the same three figures—Satan’s messenger, Adam, and Eve—appear in both P28 and its supposed models, there is no other option than to reposition Eve to the left that has been vacated through the relocation of the snake to the center. Thus, the apparent shifting of Eve to the left is part and parcel of the reinterpretation of the central figure as Satan’s agent. Although the relocation of Eve to the extreme left in P28 needs no independent explanation, the particular pose that she comes to assume must be explained on its own. Evidently, the most plausible model for Eve’s figure in P28 seems to be that of Eve in the central picture in Grandval 5v3, as identified in the traditional works (e.g., Raw 1953; Broderick 1978) and maintained in a different guise in the second version of the challenging analysis presented above, whereby the concatenation of the snake + the two figures of Eve + Adam—ranging over the two distinct pictures in the source material— would have provided a compositional basis for P28. In this view, the derivative process of the relocation of Eve from center to left periphery, contingent on the shifting of the messenger to the middle, would have largely shaped the resultant posture of Eve in P28 due to the match between the snake and the right figure of Eve in face orientation. By contrast, at first glance the derivation of Eve’s gesture in P28 poses a serious problem to the alternative account that characterizes Vivian 10v2 and Bamberg 7v2 as a prime source of inspiration for P28. Unlike the other alternative treated immediately above, no figure can be found there that even remotely resembles that of Eve eating the fruit. One must therefore appeal to the depiction of Eve comparable to that found in Grandval 5v3 as a supplementary model. This argument, however, needs qualification or even wholesale dismissal. The figure of the snake holding the fruit in its mouth and facing to the right may have largely determined the consequential pose of Eve transferred in its place. Such visual effects on the relocated figure of Eve that are assumed to have been exerted by the underlying figure of the snake in the model (Vivian/Bamberg) are in fact wholly analogous to the formative role that the original figure of Eve in Vivian/Bamberg has been claimed above to have played in determining the gesture of Satan’s agent in P28. In this light, the problem of the figuration of Eve in P28 succumbs fully to a principled account and ceases to be a challenge for the identification of Vivian and Bamberg as a primary model for P28. Furthermore, regardless of the extent to which this internally motivated account was conclusive, there is a further argument to be made for defending the postulation of Vivian/Bamberg as a primary source of inspiration in the face of the challenge made

8.2 A diachronic–comparative perspective 

 195

by the eating posture of Eve. Given that all the extant Touronian Bibles made their own distinct selection out of the stock of images available in the Cotton Genesis tradition (Kessler 1977), the use of an image that happens to be absent in some works (Vivian/ Bamberg) but was attested elsewhere in the same Touronian group (Grandval, but not San Marco; see section 12.3) would not necessarily weaken the plausibility of identifying Vivian/Bamberg as a primary model. After all, the inspiration by and borrowing from a specific model would not invariably imply that the artist copied exclusively the pictures that were attested in the source material. If the artist was familiar enough with the pictorial tradition in question, rather than passively copying the material at hand without actively and creatively putting into practice the underlying knowledge, he would have ready access to the shared system of artistic craft and knowledge that made genealogically-related, individual acts of production possible, that is, the common code of visualization that is executed with a range of variation in each concrete case. Returning to the line of specific reasoning, as Kessler (1977: 18) substantiated, the Temptation and Fall scenes realized in Vivian and Bamberg underwent abbreviation in such a way that the two separate scenes of Eve’s Picking and Eating the Fruit, as witnessed in Grandval, were integrated into one. Since there is independent evidence that the Junius artist actually used Grandval as a primary model for some Junius pictures (e.g., P10, section 6.2), his acquaintance with the cognate Grandval picture (5v3, left) seems beyond reasonable doubt. The Junius artist accordingly must have not only drawn on the actual pictures in Vivian and Bamberg resulting from a merge of the two separate scenes but also known the two original images before the conflation, as manifested in Grandval. It deserves emphasis here that the fuller representation prior to abbreviation is actually attested in one of the Touronian Bibles, Grandval. If, in the absence of Grandval, San Marco (Pl. 46) were the only testimony of the nonconflated depiction, the foregoing argument for the primacy of Vivian/Bamberg as sources for P28 would hardly prevail. Parenthetically, the foregoing argument would lend further plausibility to the inference that the Junius artist was affiliated with the Touronian School as one of its in-group members. Now that the problem of the eating figure of Eve has been accommodated successfully, it may be appropriate to discuss some of the empirical implications that can be drawn from the central figure spreading the hands to both sides in weighing the two competing accounts of P28 adduced above. This time, it is the view—including the conventional versions advocated by Raw (1953) and Broderick (1978), for example—primarily predicated on Grandval (and San Marco and Millstatt 11r [Pl.  64]) that faces criticism. From where does the figure of handing the fruit in both hands originate? A merge of the two depictions of Eve standing side by side in Grandval or San Marco, or the single figure of Eve eating the fruit, unlikely would be transfigured unequivocally into the devil’s pose in P28 as a consequence of a well-motivated derivation. The situation is worse for the conventional accounts that adhere to the single presence of Eve, for which the leftward stretching of the arm lacks demonstrable motivation. Most impor-

196 

 8 The Fall of Adam and Eve: P28/P31

tantly, the devil disguised as an angel could be inserted anywhere—at the right, the left, or the center—with equal probability. In fact, the messenger appears on the left side in P31.1. Thus, the conventional account must be rejected as least explanatory. For the view that postulates the dual presence of Eve, however, things would look more amenable to understanding, as the right and the left figure of Eve extend the outer hands to the right and to the left, respectively, from which there seems to be only a short step to reaching the devil as depicted in P28. While staying in purely formal terms of derivation, both versions appear to be equivalent, capable of accounting for the resultant central figure of Satan’s messenger in P28. Confronted with the striking resemblance between Satan’s agent in P28 and Eve in Vivian 10v2 and Bamberg 7v2, however, the two rival accounts diverge in their empirical implications that may be worth evaluation. On the one hand, the similarity is claimed to have been genealogically motivated: it would have been a consequence of borrowing and adaptation on the part of (a Touronian predecessor of) Junius 11. On the other hand, the compositional resemblance in question may have been accidental: the parallelism between P28 and Vivian/Bamberg in the arrangement of the three figures that the other account applies critical significance to would have come into being contingently parallel to but substantively unrelated to each other. That is, the similarity may well be dismissed as a pseudo issue unworthy of serious inquiry on its own. Needless to say, there is no decisive indication inherent in the phenomenon itself that may testify to the nature of the apparent analogy. Inasmuch as the first hypothesis is capable of explaining what must necessarily be left unexplained in the other, however, it may be reasonably concluded that the former should be chosen over the latter. Yet there may appear to be a possibility for compromise in the second version, corresponding to a similar auxiliary theorizing that may be added to the first version. In addition to the primary model provided by Grandval and San Marco, the segments of Vivian 10v2 and Bamberg 7v2 under discussion would have played a secondary part in shaping P28. In methodological and epistemological terms, however, the two cases of the allegedly supplementary sources are hardly equal: there is evidently an existential asymmetry in the synchrony—the amount of information—and diachrony—the evolutionary trajectory—between the group of Vivian 10v2/Bamberg 7v2 and that of Grandval 5v3/San Marco. Being abbreviations of the full version as realized in Grandval and San Marco, Vivian 10v2 and Bamberg 7v2 presuppose the double representations of Eve in both the diachronic (evolutionary) and synchronic (conceptual) dimensions: the shorter version is a derivative of the longer one, rather than vice versa. While the former owes its emergence to the latter, the converse does not hold. In the circle of artists sharing common knowledge, then, the access to the abridged version would entail familiarity with the full one. Accordingly, in the first alternative, the double depiction of Eve in general and her specific pose at the right in particular would have been part of the shared knowledge, both synchronically and diachronically presupposed underlying the use of the merged figure in Vivian and Bamberg. By contrast, superficial contact exclusively with the full version without the presupposed acquaintance with the shorter one

8.2 A diachronic–comparative perspective 

 197

cannot directly lead to a correct conceptualization of the latter. Even if a conflation of the two figures of Eve were by any chance conceived at all, the result unlikely would be determinable in unique form. Thus, as far as one adheres to Grandval 5v3 and San Marco and prioritizes them as the primary inspiration for P28, one must specifically postulate Vivian 10v2 and Bamberg 7v2 as an independently necessitated model—even if claimed to be only a secondary one—in order to account for the similarity at issue. Once the latter group has been acknowledged as a source of whatever degree of significance, this modified and compromised view comes closest to being indistinguishable from and consequently reducible to the first account exclusively based on Vivian 10v2/ Bamberg 7v2. Therefore, the determining involvement of Vivian 10v2 and Bamberg 7v2 in defining P28 must be recognized in their full capacity. A further consideration will call into question the sustainability of postulating Grandval and San Marco as primary models, thereby corroborating an unqualified recognition of Vivian and Bamberg as the major sources of inspiration and conversely prompting a total rejection of Grandval (and by implication San Marco) as immaterial. That is, the prior use of Grandval 5v3 for P20 (section 7.2) may diminish the plausibility of its alleged involvement in the construction of P28. More specifically, the Temptation of Eve in Grandval, in conjunction with the preceding scene of the Admonition, served as a model for P20. This complex source presupposed the integration of the two images in conformity with the Cotton Genesis archetype, and conversely favored separation of the left half of Grandval 5v3 into two subcomponents, the Temptation of Eve and the Fall of Adam and Eve. Such integration and segmentation of successive pictures in the Grandval registers may have deterred reintegration of the Temptation and the Fall in Grandval for additional reuse. In contrast, while picturing basically the same episodes, Vivian and Bamberg constitute distinct visualizations in which they are represented as a single scene rather than a succession of two events, as in Grandval and San Marco. This outstanding difference in pictorial arrangement would have led the artist to make a categorical distinction between the two groups of pictures and prevented the earlier use of one detracting from subsequent use of the other. In conclusion, there is every reason to assert that P28 would have been modeled on Vivian 10v2/Bamberg 7v2, and conversely there is no cogent reason to identify Grandval 5v3 as a primary source at the expense of Vivian/Bamberg. While the reconfiguration resulting in P28 induced no change in the identities of the three beings involved—Adam, Eve, and Satan’s agent—in the course of adaptation, a more drastic reconceptualization may be held responsible for the production of P31.1. Specifically, P31.1 may well be derived from the scene of the Introduction of Adam and Eve, as exemplified in all four Touronian Bibles, Vivian 10v1 (right), Grandval 5v2 (left, including the zone-dividing tree at right), Bamberg 7v2 (the second scene from the left), and San Paolo 8v2 (middle), and further at San Marco (Pl. 44). Of vital importance in the derivation just proposed is the radical conversion of God to Satan’s proxy that it is bound to postulate. As a look at the comparative materials will readily show, the rationale behind the hypothesis is the overall configurational parallel of the images concerned.

198 

 8 The Fall of Adam and Eve: P28/P31

God or Satan’s agent stands at the left just behind Eve, who faces Adam standing at the right. Both God and the emissary raise one of their hands, either touching Eve on the shoulder (Vivian, Bamberg, San Paolo, and San Marco) or not (Grandval and Junius).116 Thus, despite the change in the identities, the left figures—both God and Satan’s messenger—assume a strikingly similar posture, which can be explained by positing a derivational relation between the Touronian works (and its close relatives)—the source—and Junius 11—the goal. A notable difference between the Introduction scenes and the Temptation of Adam scene is Eve’s pose in the center. In the Introduction, Eve stands upright with her hands stretching down, whereas in P31.1 Eve, crouching and raising both hands, is humbly offering the fruit to Adam. The difference in hand gesture seems to be a relatively minor issue of adjustment that would have been executed contingent upon the adaptation of the model to the temptation scene. Unlike the scene of presenting Eve to Adam in the sources, the Temptation of Adam in the recipient manuscript requires the depiction of Eve handing the fruit to Adam as a focused act. In this connection, it may be interesting to point out that on the Hildesheim door (the second panel on the left column; Kessler 1977: fig. 5; Kaspersen 1988: fig. 1), too, Eve is raising her hands toward Adam upon encountering him (though not offering anything), a gesture that is indeed comparable to P31.1. This may suggest that the hands stretched down straight may scarcely have been an invariable distinguishing feature of the figure of Eve being presented to Adam; a variation may have been inherent in the tradition. More urgently at stake is the crouching pose of Eve, who is not represented in an analogous posture even on the Hildesheim door. Where does this posture come from, then? Is this a total innovation unprecedented in the traditional works? One need not go far in search of parallel figures; they are readily found in the Touronian Bibles. The crouching figures of Adam and Eve are attested in Vivian 10v2 (right, Reproval scene) and Bamberg 7v2 (right, Covering/Hiding scene). While appearing outside the model scene of the Introduction itself, these images occur in contiguous places—just below the Introduction and next to the model for P28 (Vivian 10v2 right), or in the same panel in which the Introduction and the Temptation of Adam and Eve are depicted (Bamberg 7v2). Also of some interest are the slightly crouching pose of Adam and Eve in Bamberg 7v3 (left, Judgment scene) and that of Adam in San Paolo 8v2 (middle, Introduction). In light of the proximity of these comparables to the primary model pictures, it seems plausible to infer that they would have provided extra inspiration for the composition of P31.1. Finally, on an additional note, the tree behind Adam may also be brought to bear on the derivational account being submitted. Closely paralleled in Grandval and San Marco, and to some extent Bamberg, the tree is depicted on the right side of P31.1. The absence of a corresponding tree in Vivian may be subject to understanding by assuming

116 Although the right hand is raised in the majority in contrast to P31.1, Bamberg is analogous to Junius with the left hand lifted.

8.2 A diachronic–comparative perspective 

 199

that Adam is depicted at the extreme right, which excludes additional room for further illustration to the right on the page. Furthermore, assuming that the San Paolo artist relocated a tree between Adam and Eve in favor of a resultant symmetrical arrangement, much as in the Temptation scene (Kessler 1977: 19), one may legitimately generalize that the presence of the tree behind Adam would have been an integral feature of the Introduction scene. Drawing on this generalization, then, it may be claimed that the corresponding feature attending on the Temptation scene in P31.1 reconfirms the proposal of deriving P31.1 from the images of the Introduction. The evolutionary trajectory from the Touronian models to the creation of P31.1 has thus been delineated in its formal aspects. It may be appropriate next to elucidate the conceptual and practical background against which the Introduction scene came to be chosen as the compositional basis for a completely different theme, the Temptation of Adam and Eve. The Introduction of Adam and Eve constitutes the beginning of human life in paradise, whereas their Temptation and Fall terminates their existence there. On both critical occasions, Eve plays a determining role in establishing their entry into and expulsion from paradise. Realizing to the full the two events to be of equally decisive significance in human history and conceptualizing them as the absolute opposites of each other would plausibly have provided a cognitive basis for making the following one—the closure—dependent on the preceding one—the opening—in pictorial representation. Thus, the whole process of adaptation here at work may be characterized as an analogical extension based on metaphoric reversal. There is still another dimension to the adaptation for reuse under consideration: practical considerations may have been involved. It must be pointed out first that the Introduction scene itself fails to be represented in Junius, not only in the extant manuscript but also, most probably, in its original state. In narrative terms, the image of the Introduction of Adam and Eve, had it really been included in Junius 11, would have occurred between P9 (Creation of Eve) and P10 (Adam and Eve before God). No traces of disruption, however, are in evidence in the current manuscript between pages 9 and 10. These two pages accordingly would have been continuous in the original state of the manuscript, as they are now. Therefore, it may be unlikely that the scene of the Introduction would have been included in the original design of the manuscript. Under such circumstances, the Introduction scene would have remained unused when the artist turned to illustrate page 31. Since almost all the materials contained in the Adam and Eve cycle in the Touronian Bibles are turned to good use for Junius 11, it would seem hard to understand that the scene in question, of all things, had been resolutely denied reuse. Equipped with the derivation of the Temptation scene (P31.1) from the Introduction as being explored here based on the formal and conceptual parallels, the apparent failure of the Introduction materials to make their way into Junius 11 may become subject to an explanatory account. Far from being rejected as useless or for whatever other reason, the Introduction scene was actually exploited for adaptation to the Germanic Genesis no less competently and effectively than were the other pictures contained in the same cycle.

200 

 8 The Fall of Adam and Eve: P28/P31

Broderick (1978: 214–215) emphatically points out that P31.2 resembles exclusively San Marco (Judgment of Adam and Eve, Pl. 50), with no other cognate images known elsewhere in the Cotton Genesis tradition. Indeed, the similarity between the two is indubitably there to be recognized, but the San Marco picture in question resembles P41.2 (Pl. 19) more to the point that the latter two might be regarded as almost identical. In both, God is situated on top, and down below are Adam and Eve at the bottom on the left and right sides, respectively. Behind them stand trees on the margins. As discussed at the end of section 8.1 above, P31.2 and P41.2 were designed as a complementary pair of cause and effect, both conceptually and compositionally; the latter would have been created under the inspiration of the former. Furthermore, as will be claimed in section 10.2 below, the group of P41.1/P44, the Judgment scenes, is a Junius innovation that subsequently found its way into San Marco (Pl. 50), largely through the integration of P41.1 and P41.2 into a single representation. Accordingly, the alleged similarity of P31.2 to the San Marco Judgment scene is based not so much on a direct inheritance from a common Cotton Genesis model (or an immediate contact between Junius and San Marco) as on a far-removed order of linkage, whereby, ultimately originating from the Covering/Hiding in the Cotton tradition, P31.2 inspired P41.2 by internal derivation, which in turn was recomposed, along with the innovative Junius Judgment scenes P41.1/P44, as constituents of the San Marco Judgment. Thus, the alleged similarity between P31.2 and the San Marco Judgment is mediated by P41.2, which stands immediately behind the San Marco scene in derivational terms.

9 The Covering of Adam and Eve: P34/P36/P39 9.1 A synchronic–structural perspective Representing Adam and Eve, face to face, covering their nakedness and deeply remorseful, P34.1/P34.2 (Pl. 16), P36.1 (Pl. 17), and P39.1/P39.2 (Pl. 18) all constitute variations of each other. All three pages involved consist of two pictures, the upper and the lower. Thus, in terms of space allocated, these variations are comparable in size as well. Although some of these pictures, for example, P34.1 and P34.2, were noted explicitly as variations (e.g., Broderick 1978: 218), others have not received due recognition. The contention being made here is that the group of pictures in question is organized as an integrated set of variants based on the inventory of parametric features, as elaborated below. Varying on the same scene, the whole spectrum of variation observed may be analyzed by and reduced to a limited set of parameters, each varying between binary specifications that stand in opposition. The postulation of particular variables is an empirical issue, of course, the legitimation of which is predicated on their overall descriptive and explanatory cogency and credibility. For now, one may posit four parameters, as substantiated in the following. First, the presence versus absence of trees is obviously open to selection. While the trees are depicted in P34.2, P39.1, and P39.2, the ground is left plain in P34.1 and P36.1. Second, the distance between Adam and Eve is subject to variation, close by versus far apart. In P34.1, P36.1, and P39.1, Adam and Eve are located in proximity, whereas they are relatively distanced in P34.2 and P39.2, with both figures aligned with the extreme left and right on the page. A third parameter to be postulated concerns whether the pair is standing or sitting. Only in P39.2 are Adam and Eve shown seated. As may be recalled, the variation between standing and sitting is in effect in some other places in Junius 11: P6 versus P7 (section 4.1) and P16 versus P17 (section 2.1). Finally, a hand gesture varies between two poses. While all examples of Adam and Eve raise one of their hands, there is a notable difference in whether the raised hand covers the face (P34.1, P34.2, P36.1, P39.2)—a gesture of grief (Broderick 1978: 217)—or points to the companion (P39.1). In addition to the four self-evident properties discussed above, Adam and Eve facing in the opposite (reflection) versus in the same (procession) direction may be brought to bear as a further pertinent parameter. Since Adam and Eve invariably face each other in the five pictures in question, the introduction of the contrary property of facing the same direction might appear unmotivated as a significant distinctive feature for descriptive and explanatory accounts of their variation. Yet allowing for a variation in face orientation will make it possible to relate P31.2 (Pl.  15; section 8.1) to the group and offer new insight into not only the synchronic organization but also the genealogical background of the array of pictures depicting the scenes of the Covering/Hiding/Reproval in Junius 11. At this juncture, the affiliation of P31.2 with the group may appear far from https://doi.org/10.1515/9783110788068-009

202 

 9 The Covering of Adam and Eve: P34/P36/P39

evident in synchronic terms, as it does not explicitly represent the covering scene that is distinctively characteristic of the five core members (P34.1, P34.2, P36.1, P39.1, and P39.2). However, the emergence of P31.2 from the Covering scene type by innovation, as assumed here, will be fully substantiated in diachronic–comparative terms in section 9.2 below. All five pairs of Adam and Eve so far discussed confront each other, with the sole exception being the new member, P31.2, in which the couple is oriented parallel in the same direction. Significantly enough, this last feature is common to Bamberg 7v2 and San Marco—two of the three closest cognates—as will be examined at length in the following section. Once P31.2 has been admitted into the set, it becomes necessary to take into account the kneeling pose of Adam and Eve relative to the features of standing and sitting. As will be seen in the following section, the distinction between standing and kneeling marks one of the distinguishing features between P41.2 (Pl. 19) and P44 (Pl. 20). In anticipation of the same differentiating pattern in evidence elsewhere, the exact same feature may be held responsible for the contrast between P31.2 on the one hand and P34.1, P34.2, P36.1, and P39.1 on the other. A question arises then about the relation of kneeling to standing and sitting, respectively. Specifically, are these three values equipollent in terms of opposition? In other words, is there a three-way distinction involved? The equipollent distinction, however, would imply that all three distinct values are more or less equal in their probability of occurrence and more or less even in their distribution among the variant members. As it is, the actual situation differs from expectation: the standing posture—four occurrences—prevails over both sitting and kneeling—each attested singly. Furthermore, while the opposition between standing and sitting (as with P6/P7)—in the absence of kneeling—and that between standing and kneeling (as with P41.2/P44)—in the absence of sitting—are in force independently, as remarked above, the contrast between sitting and kneeling in the absence of standing does not seem to be demonstrably in existence. In light of these asymmetric patterns of occurrence, one may be warranted in assuming that kneeling and sitting are covariants of each other, which are subsumed under the common variant member that constitutes the opposite of standing as an integrated unit. Put another way, standing is opposed to nonstanding at a higher level of variation, and the nonstanding variant in turn comprises two subvariants, sitting and kneeling, a variation involved at a lower level. Accordingly, the characterization of kneeling as a covariant of sitting would warrant identifying P31.2 as analogous to P39.2 in regard to their deviation from the prototype by virtue of the nonstanding pose that they commonly depict. The horizontal arrangements do not seem to serve as a significant variable, as Adam is invariably positioned on the left, or where the pair and God stand approximately on the same ground level, Adam is placed to Eve’s right, that is, closer to God, who always appear on the right in linear arrangement. Nor does the means of covering nakedness— by hand (P34.1) versus by leaves (P34.2, P36.1, P39.1, and P39.2)—seem to be encoded with parametric significance. Since the use of the leaves for covering the bodies is narratively determined (Genesis B 845a–845b “þā hīe heora līchoman / lēafum beþeahton” [Then

 203

9.1 A synchronic–structural perspective 

they covered their bodies with leaves; Anlezark 2011: 65]), the contrast can hardly be manipulated as an autonomous pictorial property open to independent control. The parametric features thus identified are distributed among the six pictures under consideration (including P31.2), as indicated in Table 10, to construct the existing variants. Table 10. Distribution of feature specifications among the variant pictures of Adam and Eve covering themselves Feature

P31.2

P34.1

P34.2

P36.1

P39.1

P39.2

With or without trees Close or far Hand raised to face or not Standing or sitting/kneeling

with close yes/no kneeling

without close yes standing

with far yes standing

without close yes standing

with close no standing

with far yes sitting

Adam and Eve facing each other

no

yes

yes

yes

yes

yes

A cursory look at the overall distribution of the parametric features indicates that the four pictures excluding the two on the extreme right and left columns—P31.2 and P39.2—are minimally differentiated, that is, they are distinguished by a single parameter. Therefore, the middle four pictures resemble each other maximally. By contrast, the remaining two differ from each other as well as from the other four along multiple parameters, with the single exception being the pair of P34.2 and P39.2, which is distinguished solely on the basis of standing versus sitting. Maximally differentiated from the remainder is P31.2, which is unique in depicting Adam and Eve oriented in the same direction. The distribution pattern of these features may translate readily into the perceptual judgment that the middle four instances are more similar to each other than to the other two, which seem rather isolated on account of their overall lesser resemblance to the other group members, except for the privileged pair (P34.2/P39.2) noted above. Because of the minimal distinction involved, P34.2 and P39.2 may be characterized precisely as standing and sitting variants of each other. Excluding for the moment P31.2—the most peripheral member of the group by virtue of its unique feature of Adam and Eve not facing each other (Table 10)—and thus focusing on the strict group of the Covering scene displaying Adam and Eve oriented face to face, in purely logical terms the selection of each parametric value on an independent basis may amount to maximally sixteen differing combinations (24), of which five are materialized. Actually, however, cooccurrence of parametric values is constrained in large measure, as substantiated below, and accordingly not all logical combinations of features are expected to occur with equal probability. Furthermore, the selection of these variable features partly depends on the vertical arrangement of pictures on the page in which they are used. Specifically, the absence of trees is limited

204 

 9 The Covering of Adam and Eve: P34/P36/P39

to the upper pictures (P34.1 and P36.1), thereby entailing that the lower ones are provided with trees, as borne out by P34.2 and P39.2. Moreover, the parameter of close by/ far apart is correlated to the location of the picture on the page. All three top pictures (P34.1, P36.1, and P39.1) show Adam and Eve standing close by, whereas both of the bottom pictures (P34.2 and P39.2) situate the pair far apart from each other. In light of the actual distribution of parametric values represented in the above table, one might be induced to generalize the following relation of markedness, a relative valuation along the scale for conceptual prototypicality and likeliness of occurrence: the more marked a feature is, the less typical and the more deviant from the prototype it is perceived to be and the less likely it should occur. The unmarked (or default) values are encoded in the following features in contrast to their opposites: standing (4/5; four occurrences of the total five) against sitting/kneeling, hand to face (4/5) against hand pointing to the companion, close by (3/5) against far apart, and with trees (3/5) against without trees. Yet the small sample size makes it infeasible to detect any statistical difference among varying combinations of these features. Of no less interest, all five manifestations exhibit at least one marked value according to the criteria just presented. That is, there is no realization in which all four parameters concerned are unmarked as to their specification. Concretely put, there is no picture in existence in which Adam and Eve stand close to each other, raising one of their hands to cover their own faces, and with large trees growing on the ground. Conversely, no picture is attested in which Adam and Eve are seated or kneeling apart from each other on the bare ground, raising their hands not to cover their faces but to point to the companion. Continuing the examination of the Covering group in the strict sense, it deserves particular attention that only P39.2 exhibits two marked features—sitting and located at a distance. Furthermore, among the other four members with a single marked value, P39.1 is differentiated from the others by virtue of bearing the unique property of not raising one hand to the face. This image is comparable to its lower half, P39.2, in the sense of carrying a unique feature value: in addition to the marked specification of the proximity feature, P39.2 is uniquely encoded with the “sitting” value. Thus, both pictures on page 39, to the exclusion of others, are assigned a unique specification in regard to one of the four parameters involved. The remaining marked specifications— far apart and without trees—are shared by two members. Given that both constituent pictures are uniquely specified, it may be legitimate to assume that P39 would have been distinctly of a derivative status relative to the other two pages. In this connection, the occurrence of P39 subsequent to the two others—P34 and P36—seems of paramount interest. More concretely, one may generalize that the further the artist proceeded in composing variants of the same type, the further the products were differentiated from the prototypical instance for the sake of variation. It must be pointed out in this connection that P34.1, the initial instance of the series, hardly can be counted as prototypical, because Adam and Eve do not cover their nakedness with leaves, a typical means of hiding in the traditional images of Covering/ Hiding, as confirmed in all relevant exemplifications, especially the Hortus Deliciarum,

9.1 A synchronic–structural perspective 

 205

Bamberg, and San Marco, as discussed in the following section. Given the prototypicality thus determined to be inherent in P34.2, this picture should be identified as the Junius instance that is closest to the cognate representations of Covering/Hiding. In this light, the generalization presented above needs expansion, as one must now take into account the backward as well as forward direction. The further a variant is distanced in its location from the prototypical realization, the more deviant it becomes in its manifestation. This revised generalization bears empirical verification. Standing at an equal distance from the pivotal point (P34.2) in both directions, P34.1 and P36.1 deviate from the prototypical instantiation exactly to the same extent, that is, by the marked value of “without trees.” By contrast, P36.1 and P39.1 on the one hand and P39.1 and P39.2 on the other bear out the differing degrees of deviation in accordance with the generalization. P39.1 may count as less typical because the negative value for the feature of “hand raised to face” responsible for the definition of P39.1 is rarer in manifestation than “without trees” characteristic of P36.1 (and P34.1). By the same token, P39.2, farthest distanced from the prototype, is correspondingly maximally removed from it in manifestation on account of the two marked specifications involved. In the backward direction, a correspondingly graded differentiation obtains as well, if, as suggested above, P31.2 is integrated into the group of pictures at issue under the assumption that kneeling constitutes a variation of “sitting” and that as such it embodies a marked value against “standing.” Because of its unique occurrence among the group, the feature of “sitting” may be regarded as more marked than the marked values of “far apart” and “without trees,” which are embodied by more than one instance, respectively. Therefore, the feature of “kneeling,” a variation of “sitting,” may be assigned a degree of markedness comparable to that of “sitting.” Moreover, only Eve raises her hands to her face in P31.2 (hence yes/no in Table 10), thus deviating still further from the normal composition characteristic of the group in which both Adam and Eve make the same gesture. Therefore, there is every reason to determine P31.2 as more removed than P34.1 from their prototypical variant (P34.2). Up to this point, I have been concerned with the inner organization of the group of pictures representing the Covering/Lamenting scenes. The following investigation addresses the external dimension of the group’s organization, that is, its relation to other pictures in the same poem. Excluding the most peripheral member, P31.2, the remainder—P34.1, P34.2, P36.1, P39.1, and P39.2—constitute a homogeneous group of the Covering/Lamenting scenes. As such, the set may be regarded as being in opposition to that of P10, P11, and P13, which all represent Adam and Eve in paradise, the state of affairs that arises upon introduction of Adam and Eve, namely, the beginning of human life in paradise. By contrast, the situation depicted in the opposition group of P34.1, P34.2, P36.1, P39.1, and P39.2 precedes the ending of human life in paradise. In terms of conceptual polarity, rather than temporal sequencing of events, however, the diametrical opposite of P10/P11/P13—God’s Blessing—should rather be identified with P41/P44/P45.1—God’s Judgment. Correspondingly, the set of P34/P36.1/P39—representing the situation that obtains immediately before Adam and Eve are confronted with God’s Judgment—may

206 

 9 The Covering of Adam and Eve: P34/P36/P39

well be characterized as subvariants of P41.2/P44/P45.1, and should accordingly be subsumed under a common superordinate group, together with P41.2/P44/P45.1. In this earlier stage preceding God’s Judgment, the couple Adam and Eve, alone to themselves, confront each other in nakedness and remorse. Thus, it is in conjunction with P41.2/P44/ P45.1 that P34.1/P34.2/P36.1/P39.1/P39.2 are perceived to be in opposition to P10/P11/P13. The primary variants of this complex group (consisting of two subgroups), therefore, are represented by P41.2/P44/P45.1, as they are maximally distinguished from P10/P11/P13 in conceptual terms, as noted above. Given the precise determination of the structural–conceptual relationship in which the three groups of pictures are situated as presented above, it may now be appropriate to elucidate the substantive basis of their opposition and variation by identifying a set of distinctive features underlying their organization. The following observations on intergroup variation patterns may be of significance in specifying relevant properties for contrast and variation between the groups. First, while the group of P34.1, P34.2, P36.1, P39.1, and P39.2 is set out in the field, P10, P11, and P13 are all enclosed in an architectural frame. The enclosure in the architectural structure implies full protection by God, whereas the absence of the frame hints at vulnerability to all danger. Second, no animals appear in P34.1, P34.2, P36.1, P39.1, and P39.2, whereas the group of P10, P11, and P13 contains animals in all three member pictures. Third, God is invariably depicted in P10/ P11/P13, but he is always absent in P34.1, P34.2, P36.1, P39.1, and P39.2. Fourth, P34.1, P34.2, P36.1, P39.1, and P39.2 each occupy an upper or lower half of a bipartite illustration. By contrast, P10, P11, and P13 constitute autonomous illustrations standing alone. Fifth, as pointed out above, P34.1, P34.2, P36.1, P39.1, and P39.2 all show Adam and Eve facing each other, but the opposite group of P10/P11/P13 is varied along this feature, as in one of them—P10—Adam and Eve are oriented in the same direction. Sixth, while the variation between standing and sitting holds in the set of P34.1/P34.2/P36.1/P39.1/P39.2, as discussed above, they are consistently shown standing in P10, P11, and P13. Seventh, while the group comprising P34.1/P34.2/P36.1/P39.1/P39.2 is again variable in regard to the distance between the pair, they are never set apart in the latter group of P10/P11/P13. Of particular interest in this connection is a contrast between P34.2 and P11, which would otherwise be almost identical, with a focus on the representation of the pair. Indeed, in both pictures, the pair standing face to face are separated by a tree in the center; in this sense, Adam and Eve are not in immediate proximity. In P34.2, however, two additional trees intervene between Adam and Eve, whereas the two figures are placed directly beside the central tree without a gap in P11. Instead, the two additional trees are placed one behind Adam and one behind Eve. This difference in spacing and arrangement seems to be of vital importance, lending substantive support to the analysis whereby P11 shows the pair standing close by, much as do P10 and P13 unequivocally. Of particular interest in this regard is P13, in which the two figures standing next to each other are intersected by a tree and accompanied by another tree behind each, a composition conspicuously analogous to P11. Thus, much as Adam and Eve are indisputably adjacent in P13, the couple should be viewed in parallel fashion in P11.

9.1 A synchronic–structural perspective 

 207

Eighth, the appearance of trees is optional in both groups, although they are depicted more often than not. For a better understanding of the overall organization, it will be instructive to compare the group of P34.1/P34.2/P36.1/P39.1/P39.2 with the set of P41.2/P44/P45.1, which is a complement subgroup to it, and, put together, is opposed to the class of P10/ P11/P13 at a higher level of organization. First, a setting in an architectural frame— present in P41.2 and P45.1 but absent in P44—is optional in contrast to both the opposite group of P10/P11/P13—always represented—and the subgroup of P34.1/P34.2/P36.1/ P39.1/P39.2—never represented. Second, animals are nowhere to be seen like the companion subgroup and in distinction from P10/P11/P13. Third, while totally absent in P34.1/P34.2/P36.1/P39.1/P39.2, God is always represented in P41.2/P44/P45.1 as in P10/ P11/P13. Fourth, P41.2/P44/P45.1 are varied in the mode of their existence: while P41.2 and P45.1 are constituents of larger pictures, P44 exists on its own as a full-page illustration. This mixed character differs from both P10/P11/P13—always standing alone—and from P34.1/P34.2/P36.1/P39.1/P39.2—always in combination with the other half. Fifth, Adam and Eve are arranged symmetrically (P44) or in a uniform direction (P41.2 and P45.1), much as in the contrary group of P10/P11/P13 and in contrast to P34.1/P34.2/ P36.1/P39.1/P39.2, in which the face-to-face orientation invariably obtains. Sixth, much as with the other subgroup, the feature of standing is variable, as Adam and Eve are standing (P44 and P45.1) or kneeling (P41.2). This is in contrast to the invariably standing pose in P10/P11/P13. Seventh, the relative distance between Adam and Eve varies between two values, close by (P45.1) and far apart (P41.2/P44), in parallel to the other subgroup, whereas the opposite group of P10/P11/P13 shows the pair standing closely together. Eighth, the treatment of trees proves to be indistinguishable among the three sets, as the trees are optional elements throughout. By way of providing an overview, Table 11 indicates the differing distributions of parametric values in the three groups under examination. Table 11. Distribution of feature specifications in the three groups of pictures Feature

P10/P11/P13

P34.1/P34.2/P36.1 /P39.1/P39.2

P41.2/P44/P45.1

Architectural frame Animals God Stand-alone/combined Pair’s position

yes yes yes stand-alone facing the same way/ face to face standing close yes/no

no no no combined face to face

yes (simple)/no no yes stand-alone/combined facing the same way/face to face standing/kneeling close/apart yes/no

Standing/sitting Close/apart Trees

standing/sitting close/apart yes/no

208 

 9 The Covering of Adam and Eve: P34/P36/P39

The subgroup P34.1/P34.2/P36.1/P39.1/P39.2 is often differentiated from the other subgroup P41.2/P44/P45.1 by maximizing differences from the opposite group P10/P11/P13. Specifically, an architectural frame is never depicted, God is always absent, there is no full-page picture, and Adam and Eve are invariably represented face to face (Table 11). The ways in which these elaborated differentiations are devised and implemented are comprehensible, given that the subgroup of P34.1/P34.2/P36.1/P39.1/P39.2 is derivative and marked, being secondary to its primary counterpart (P41.2/P44/P45.1), and correspondingly removed from the major opposite group P10/P11/P13 maximally in structural terms. Thus, the maximal differentiations applied may be regarded as effective visualizations of the underlying conceptual/compositional network of the three (sub) groups of images at issue.

9.2 A diachronic–comparative perspective From a comparative perspective, P31.2 exhibits an outstanding similarity to Bamberg 7v2 (right; Covering/Hiding) in specific details: in both, only Eve raises her hand(s) to her face. The couple Adam and Eve face in the same direction rather than confronting each other. Apparently, however, there are a number of notable differences between the two. First, the orientation of the figures differs, leftward looking (P31.2) on the one hand, and rightward looking (Bamberg) on the other. This alteration may readily be attributed to the presence of the devil at the right in P31.2, because an adherence to the rightward orientation would possibly result in a contradictory gesture of Adam and Eve praying to the devil. Moreover, the positioning of the devil at the right is independently motivated, as elucidated in section 8.1 above. For more on the genealogy of Satan’s emissary, see toward the end of this section. Second, while Adam and Eve do not cover their nakedness in P31.2, they do so emphatically in Bamberg 7v2. This difference, however, may be explained by hypothesizing the split of originally a single scene of Covering into Praying and Covering in Junius, an innovation that will be justified substantially below. With the Covering component separated out and assigned exclusively to the following pictures (P34.1 and P34.2), P31.2 is dissociated from this component of signification and becomes free not to represent a Covering act. Third, a tree is shown growing between Adam and Eve in P31.2, whereas one stands behind the couple in Bamberg. This mismatch in the location of the tree is no less amenable to explanation in terms of derivation. The reversal in body orientation—leftward versus rightward—implemented in Junius 11 as pointed out above requires in turn a relocation of the tree. Otherwise, the pair would look like they were praying to the tree. Notice in this connection that, despite the chain of relocations executed, the presence of the tree behind Adam is retained unchanged, common to both the model and its derivative. Now that all the major distinctions are subject to credible explanations based on a trajectory of derivation from one to the other, it may legitimately be claimed that

9.2 A diachronic–comparative perspective 

 209

P31.2 and Bamberg 7v2 are traced back to a common source, and more specifically, that the former derives from the latter, or put more cautiously, it evolves from a model that seems to be closely reflected in Bamberg 7v2. As substantiated at length above, P34.2 is characterized as the prototype of the series depicting the Covering/Lamenting. Searching beyond Junius 11 for parallels to this exemplary instance, one may recognize comparable instantiations in the Hortus Deliciarum 17v (Pl. 69, upper right; Green et al. 1979: pl. 11, no. 20), Bamberg 7v2 (right, Pl. 23), and San Marco (Pl. 47 [Covering], Pl. 48 [Hiding]). Common to all three as well as to P34.2, the couple is apparently hiding under the tree in the absence of God. As commented by Broderick (1978: 217), the Hortus Deliciarum seems by far the closest image of the three in detail to the Junius picture. Adam and Eve, represented in mirror images, face each other, raise one of their hands to their faces in a gesture of lamentation, and cover their nakedness with leaves. In the remaining two pictures, by contrast, Adam and Eve are shown oriented in the same direction rather than confronting each other. Only one of the couple is raising one hand, either to reach the face (Eve; Bamberg) or to point forward away from the face (Adam; San Marco). Despite the remarkable similarities in evidence between P34.2 and the Hortus Deliciarum, however, the linear arrangement of the pair is reversed, with Adam standing at the right in the Hortus Deliciarum. Furthermore, while Adam and Eve stand far apart with three trees intervening in P34.2, these two human figures are close to each other in their position in the Hortus Deliciarum, with only a single tree—the comparatively thin trunk—standing between them, far from overgrown with leaves in contrast to those depicted in P34.2, Bamberg, and San Marco. Noteworthy in this connection is the absence of corresponding images of the Covering/Hiding scenes on their own in Vivian, Grandval, and San Paolo, although the three works, especially Vivian, would have served demonstrably as primary models for Junius 11 in many other visualizations. While these works certainly do not contain a picture in which Adam and Eve are seen alone and covering their nakedness, the comparably crouching and covering couples are shown standing side by side in front of God in Vivian 10v2 (right) and Grandval 5v3 (right)—mirror imaged to each other—in corresponding yet distinct ways (Kessler 1977: 19–20).117 In both, God raises his arm in reproval of the pair, who, however, react differently in the two works. In Vivian, Adam and Eve keep shrinking in fear, but in Grandval, while huddled up, they make excuses and deny the blame, Adam pointing to Eve, and Eve in turn pointing to the serpent, to evade reproval. The Denial of Blame is independently represented in Bamberg (7v3, left; Kessler 1977: 20) and at least in part at San Marco (Pl. 49), where only Adam denies the blame by pointing to Eve as responsible. Of particular interest here is Bamberg, which represents the Covering/Hiding scene on its own, in separation from the Denial of Blame. This autono-

117 The Hildesheim door shows a composition similar to Grandval, with a difference that the tree stands between the couple rather than between God and the couple as in Grandval.

210 

 9 The Covering of Adam and Eve: P34/P36/P39

mous Covering/Hiding scene in Bamberg proves to be strikingly similar to the couple’s huddling together in the Reproval scene in Vivian 10v2, with Eve raising one hand to the face. In light of this configurational parallelism to Adam and Eve’s pose in the Covering/ Hiding scene in Bamberg, and in the absence of a Hiding image proper in Vivian, it may be legitimate to infer that the two distinct scenes reflected in Bamberg would have been combined into one through conflation in Vivian, presumably to save space and/or to simplify the narrative development for effective presentation (cf. Kessler 1977: 20). Of further interest in this connection is the Reproval of Adam and Eve at San Marco (Pl. 49), especially Eve’s gesture of raising one hand in front of her face. As may readily be recalled, Eve does not explicitly deny blame, as does Adam, at San Marco, in distinction from Grandval and Bamberg. By contrast, neither in the Covering scene nor in the Hiding one at San Marco is Eve shown raising a hand in front of her face. That is, while San Marco makes a maximal distinction among the Covering (Pl. 47), the Hiding (Pl. 48), and the Reproval (partly instantiating the Denial of Blame concomitantly; Pl. 49), the huddling pose constitutes a privileged representation in the last scene, in contrast to Bamberg, in which the corresponding posture is found common to the Covering/Hiding and the Reproval/Denial. Also of interest is a radically condensed rearrangement of San Paolo 8v2 (right, Pl. 34). Covering themselves with leaves, Adam and Eve stand face to face with a tree growing in between around which the snake entwines. Adam raises one hand with a gesture of eating, and Eve, looking the snake in the eye, extends her hand to Adam, as if to give him the fruit. Immediately behind Eve stands God, who raises his arm high as a sign of reproval. This compositionally complex image would have resulted as a consequence of merging a series of originally separate scenes, the Temptation and Fall, the Covering/Hiding, and the Reproval (possibly including the Denial of Blame, if Adam’s and Eve’s raised hands were interpreted at the same time as gestures for identifying who is to blame) into a single representation, as interpreted by Gaehde (1963: 146–148; see also Kessler 1977: 19). Despite the far-reaching merger, the resultant figures of Adam and Eve appear to be akin to those in P34.2, with the snake and God disregarded from comparison. Similar to San Paolo is the Millstatt Genesis. In it, the Fall scene (fol. 11r, Pl.  64), which is independent of the immediately preceding Temptation (fol. 10r, Pl. 63), anticipates the covering of their nakedness. Thus, Millstatt comes close to San Paolo in representing Adam and Eve in a facing position separated by a tree, eating the fruit and covering their nakedness at the same time. Naturally, however, there is no depiction of a snake around the central tree, as the Temptation is illustrated elsewhere (10r, Pl. 63). In any event, Green’s (1955: 345) characterization is fully acceptable that the Millstatt picture owes to a conflation of the Fall and the Covering/Hiding (cf. Kessler 1977: 19). Summarizing the foregoing comparative examination, the crouching and covering figures of Adam and Eve, whether or not accompanied by a gesture of raising one hand to the face or its vicinity, appear in the Covering/Hiding scene in Bamberg, the Hortus Deliciarum (not crouching), and San Marco (not definitely crouching), on the

9.2 A diachronic–comparative perspective 

 211

one hand, and the Reproval/Denial scene in Vivian, Grandval (not definitely crouching), Hildesheim, and San Marco, on the other. Additionally, the corresponding couple is incorporated into the multiple conflation of the Temptation/Fall, the Covering/Hiding, and the Reproval/Denial in San Paolo. In terms of configuration, one may recognize two variant representations of the Covering/Hiding figures, primarily with respect to the face orientation of Adam and Eve, the variation that crisscrosses the difference in the framing context in which they are embedded. The facing couple of Adam and Eve covering their nakedness, with a tree standing between, is attested in the Hortus Deliciarum (the Covering/Hiding, the hand raised to the face exactly as in P34.2), San Paolo (through conflation; see above), and Millstatt (through conflation; see above). By contrast, Adam and Eve covering themselves but facing in the same direction are depicted in Bamberg (Covering/Hiding), Vivian (Reproval), Grandval (Reproval/Denial), Hildesheim (Reproval/Denial), and San Marco (on the three separate scenes—the Covering, the Hiding, and the Reproval/Denial). These two varying configurations of Adam and Eve, then, may prompt the following question: Which composition is closer to the original? According to Gaehde (1963: 147; 1971: 368) on the derivation of the San Paolo instance, followed by Kessler (1977: 19), for example, the San Paolo artist, in redesigning for the conflated depiction he came up with, may have adopted a symmetrical arrangement from the Fall scene commonly executed in the Cotton Genesis tradition and beyond. If this argument is correct, the symmetrical configuration in general of Adam and Eve covering themselves—the first of the two variations identified above—would correspondingly have been secondary in development to the asymmetrical version. In more substantive terms, in two of the three extant Covering/Hiding scenes standing on their own (Bamberg and San Marco), Adam and Eve are oriented in the same direction even in the absence of God, that is, in situations where the unidirectionality involved cannot be attributed to a canonical orientation to the residing God. Perhaps, the couple looks toward, or conversely away from God, as he is anticipated to make his appearance at any time. In the remaining instance (Hortus Deliciarum), the symmetrical arrangement of the covering figures is strikingly parallel to the Temptation/Fall immediately to the left in the same panel of illustration (Pl. 69; Green et al. 1979: pl. 11, no. 19). Furthermore, the placement of Adam to the right of Eve in the Hortus Deliciarum is contrary to the arrangement in Bamberg and San Marco with Adam allocated to the left. Thus, the Hortus Deliciarum image is doubly divergent from its counterparts in Bamberg and San Marco: it is symmetrical, and Adam stands at the right. While the first distinction may be ascribed to the secondary readjustment aiming at a symmetrical composition as noted above, the latter difference may be understood in comparable fashion as a consequence of an analogical redesigning. Interestingly enough, Adam occupies the same position relative to Eve in the adjacent Temptation/Fall picture as well (Pl. 69). This allocation, however, is far from peculiar, given that the identical positioning of Adam at the right is obtained in many instantiations of the same scene of the Temptation/Fall, as in Bamberg, Grandval, Vivian, and San Marco. Thus, of the two analogous

212 

 9 The Covering of Adam and Eve: P34/P36/P39

pictures—Adam (right) and Eve (left) facing each other with a tree in between—placed side by side in the same panel in the Hortus Deliciarum, it is that of the Covering found on the right side that deviates from the expected—from a comparative perspective— patterns along a cluster of variables. Hence, it may reasonably follow that the right member of the paired pictures would have been modified through assimilation to the left one by virtue of the latter’s more prominent profile in the tradition and with a view to achieving greater harmony between the two constituent pictures of the same panel. It deserves emphasis here that the assimilatory alteration invoked for reversing the linear order of Adam and Eve for the Hortus Deliciarum would also have borne directly on the symmetrical rearrangement of the covering Adam and Eve in some Cotton Genesis members including Junius 11 (see the following paragraphs). In this way, the symmetrical positioning of Adam and Eve in the Hortus Deliciarum Covering, in conjunction with their relative location, may be demonstrably attributed to a secondary process of recomposition whose model is immediately available on the same manuscript page. The above arguments have converged to indicate that an earlier representation of Adam and Eve covering themselves is reflected in Bamberg and San Marco, in which the couple, alone by themselves, are covering their nakedness with a crouching pose (notably in Bamberg) and looking in the same direction, rather than face to face. From this group, a derivative one, more frequent in attestation, would have developed through a symmetrical rearrangement of Adam and Eve, probably on the model of the Fall scene. Since this reconfiguration in no small measure would have been promoted by a general predilection for symmetry, it may presumably have been implemented on multiple occasions independent of each other. Accordingly, as Kessler (1977: 19) remarked specifically in connection with a compositional similarity between Millstatt and San Paolo, the resulting resemblances among the secondary group (section 1.2) would not necessarily have been due to a common innovation. These comparative findings have in turn significant implications for determining P31.2 and P34.2 and their mutual relationship in more precise terms. While P34.2, the prototypical representative of the Covering in Junius, is affiliated with the later group embodying the symmetrical arrangement of the couple, P31.2 exhibits notable resemblances to Bamberg insofar as the configuration of Adam and Eve is concerned, as pointed out at the beginning of this section. Since the tempter of Adam and Eve disguised as an angel is obviously a narrative innovation in the Old English/Old Saxon Genesis, one may well leave it out of account—for the time being—in comparing with the traditional biblical recensions. In essence, then, the crouching pose (as reflected in Bamberg most clearly) is exaggerated to result in a kneeling posture; and the raising of one hand to the face (as with Eve in Bamberg) is generalized to involve both hands and Adam as well. By contrast, P34.2 arises through a symmetrical rearrangement of the pair and a generalization of the raising hand gesture to both of the pair, as in the Hortus Deliciarum. Consequently, with a reference point determined at P34.2, which inherits most of the model features, the preceding two images and the following three (excluding P36.2, which is extraneous to the series of Adam and Eve Covering, Hiding,

9.2 A diachronic–comparative perspective 

 213

and Reproved; Table 10) are subject to differing degrees of variation in proportion to their distance from the focal picture, P34.2, as detailed in the preceding section.118 Between P31.2 and P34.2—the exemplary instances of the two scenes now categorically distinguished—stands P34.1 in transition. This picture is transitional not only in a locational sense but also, more importantly, in visualization. In it, the two figures cover themselves with their hands, not the leaves, as in the subsequent pictures. Inasmuch as the Covering prototypically involves leaves as a means of concealing, P34.1 falls short of this defining property. Nonetheless, the covering act itself is depicted, thus definitely subsumable under the Covering scene type. Accordingly, it may be generalized that P34.1 mediates between P31.2 and P34.2 in figural and sequential terms: compositionally and iconographically, it is closer to P34.2, but staying on the periphery of the group that is headed by P34.2, it intersects with P31.2. Contrary to the proposed derivation from Bamberg, P31.2 appears to have no bearing at all on the Covering scene in purely synchronic terms, as all it depicts is Adam and Eve praying and/or grieving without the slightest hint of covering themselves. In terms of narrative contents, the illustration corresponds very well to Genesis B 765b–783a, in which Adam and Eve, stricken by sorrow, remorse, and fear, kneel in prayer. Of immediate relevance is the following passage (777b–781a): “hwīlum tō gebede fēollon // sinhīwan somed / and sigedrihten // gōdne grētton / and god nemdon, // heofones waldend” (At times the wedded couple fell down in prayer, and spoke to the good victorious Lord and called on God, the ruler of heaven; Anlezark’s [2011: 59] translation). No mention is made of the couple covering their nakedness at this point,119 although they are said to have recognized their nakedness at 783b–784a: “bare hīe gesāwon // heora līchaman” (They noticed their naked bodies; Anlezark’s [2011: 61] translation). Diachronically and compositionally viewed, however, P31.2 should be identified as a reflection of Bamberg 7v2, an autonomous, exemplary, and earlier representation of the Covering/Hiding, as argued above. This apparent disparity between the two characterizations—art historical versus textual, each motivated on its own credible grounds— may be fully harmonized by reconstructing a conceptual split of the originally single scene of the Covering/Hiding—as materialized in Bamberg 7v2 in its older form—into the Praying/Grieving (P31.2) and the Covering (P34.1 and thereafter), and still further the Calling/Hiding (P41.2, Pl. 19; see section 10.1). This conceptual division has its own inherent moments for reorganization in the inherited convention: as substantiated above, the same scene of the Covering is represented by two distinct variants, which are primarily distinguished by the parameter of the orientation of Adam and Eve, or put another way, the contrast between the symmetrical and asymmetrical arrangement of 118 Or, less likely, just a single preceding image (P34.1), if one could no longer afford to appreciate a synchronically meaningful relation at all between P31.2 (Praying) and P34.2 (Covering/Hiding); on this issue, see the following paragraphs. 119 Only subsequently at 845a–845b “þā hīe heora līchoman / lēafum beþeahton” (Then they covered their bodies with leaves; Anlezark’s [2011: 65] translation) are they said to cover themselves with leaves.

214 

 9 The Covering of Adam and Eve: P34/P36/P39

the couple. Drawing on the conventional variation on the Covering that was available for use in the Cotton tradition, the Junius artist, then, would have forged a conceptual distinction between these formally differing but functionally equivalent variants. As a consequence of this conceptual innovation and elaboration, the two variant representations of the Covering/Hiding came to be redefined as contrastive pictorial units that were encoded with distinct meanings, the Praying/Grieving on the one hand and the Covering on the other. This innovation of constructing a distinct image for the Praying/Grieving scene independent of the Covering was motivated ultimately on textual grounds. It was a pictorial response to the novel narrative material composed in Old Saxon in which immediately following the Fall and up to God’s return to paradise, Adam and Eve are described at length praying to God and talking to each other in grief, fear, and remorse. Thus, challenged by the need to provide a pictorial representation of the praying and grieving Adam and Eve, the Junius artist reorganized the traditional formal variation without functional distinction that governed the inherited practice of picturing the Covering scene, and settled on the older, asymmetrical variant as the choice. Subsequently, the artist would have had to make minor compositional alterations to achieve better results, as remarked earlier in this section. Yet a question still remains to be answered: Why was the earlier variant chosen as the image for the Praying/Grieving scene rather than the derivative one? It should not be difficult to understand the selection. The decision would have been based on the then current practice of using the newer variant—with the symmetrical arrangement— with greater frequency for the Covering on its own or in conjunction with other scenes through conflation, as observed above. Simply put, the higher demand for use would have attached a stronger association with the Covering scene to the symmetrical variant rather than the competing one, thereby leading the Junius artist to take for granted this prevailing variety as an optimal pictorial vehicle for the Covering proper and appropriate the remaining minor one for reuse. The derivation of P31.2 from Bamberg 7v2 thus claimed has its credibility reinforced by further arguments. Of paramount interest is the depiction of the devil, notably its pose and position. Bringing in the figure once excluded from examination above, Satan’s gesture of glancing back and raising both arms in P28 would have been largely a redesigning of the figure of Eve in the Temptation and Fall scene, manifested in Bamberg and Vivian, as substantiated in depth in section 8.2 above. It can hardly be a coincidence that the very same figure appears in the panel next to that of the covering Adam and Eve in Bamberg 7v2. In light of the immediacy in space occupation, it would be quite likely that, while drawing on the crouching figures of Adam and Eve as primary material for reconfiguring them as the praying and grieving ones, the artist would have taken additional notice of the figure of Eve spreading her arms high—exactly the model for representing the devil in P28—standing to the left at the center of the adjacent panel. Since the same being—Satan’s messenger—continues to play a decisive role in the scene immediately following the Fall as in the Temptation, and since similar significations—as argued in

9.2 A diachronic–comparative perspective 

 215

the preceding section—may be inscribed in the analogous figurations involved, it would be hard to deny the conjecture that the illustrator would have been intrigued to reuse the figure of Eve in question, in line with the Temptation/Fall, for the directly succeeding scene as well. Of specific significance here is the location of Satan’s emissary: in P31.2, he stands on a hilly ground, slightly distanced from the couple, a composition that is strikingly parallel to Bamberg 7v2. The presence of the devil in P31.2 does not simply strengthen the claim that Adam and Eve praying comes from the covering couple, as reflected in Bamberg 7v2 (right). It has more specific implications by indicating that the actual model would likely have been provided by the Bamberg scene as it actually was situated as part of the extant full-page, frieze-based illustration, rather than an abstract scene type of the Covering in general, of which the image in Bamberg 7v2 is a particular instantiation. Since the presumable inspiration would have been drawn from the figure of Eve appearing in a separate panel, the Covering scene conceptualized in the abstract would unlikely have been associated with this critical source of the devil on inherent grounds. Only by reference to the adjacent picture in Bamberg 7v2 beyond the boundary of individual scenes would the vital conceptual link between the two independent images standing side by side have been activated as a source of inspiration. The replication of the figure standing on a hillock at a distance thus proves to be of overwhelming importance in specifying the probable source of P31.2. The scene of Adam and Eve sitting apart is unknown in the Cotton Genesis family, as pointed out in Raw (1953: 53–54) and Broderick (1978: 223–225). As argued from a synchronic perspective in the preceding section, it may be reasonable to infer on formal grounds that the sitting scene emerged as a variation of the normal standing one. That is, P39.2 is a sitting version of P34.2, constructed through multiple marked specifications (including the value of sitting as opposed to that of standing) and defined accordingly as the farthest removed from the prototype (P34.2). If this synchronic–structural analysis is valid, one would hardly be required to postulate the Eastern Octateuch tradition as a formative force behind Junius 11, despite Raw’s (1953: 54) contention and Broderick’s (1978: 223–225) suggestion to the contrary. The apparently unique sitting scene may accordingly be characterized as an innovation that was accomplished on a purely internal basis. One notable advantage of this alternative account is that it can readily eliminate the need to invoke the Greek Octateuchs otherwise unmotivated in the picture series under discussion (cf. Raw 1953: 54).

10 The Judgment of Adam and Eve: P41/P44/P45.1 10.1 A synchronic–structural perspective P41 (Pl. 19) consists of two pictures, one (P41.1) on top of the other (P41.2). Below, P41.2 may be regarded primarily as the scene of God calling Adam and Eve, who are hiding from him, as is explicitly interpreted by Ohlgren (1992: 92). The large spreading branches that may keep the couple from sight seem to highlight the hiding act. The succession of acts following the scene of God addressing the couple, namely, the Reproval of God and the Denial of Blame, are not specifically depicted here or elsewhere in Junius, in contrast to Bamberg, Grandval, and Hildesheim on the one hand—Reproval and Denial—and to Vivian and San Paolo on the other—Reproval only (Table 12; for details, see section 10.2). The lack of the denying act might make a pairing with the picture above (P41.1) less than optimal. In it, God judges the snake standing upright, which, upon being sentenced, slinks away on the ground. On the face of it, this judgment scene should most naturally be combined with the Judgment of Adam and Eve in P44 (Pl. 20). In line with the Denial of Blame as represented in other Cotton Genesis works, God would then have been shown judging the three beings individually in turn: representing the triad in a single page would have been logically consistent and most effective in presentation, as at San Marco (Pl. 50) at a superficial level of reading. The avoidance of this triple representation in favor of coupling P41.1 and P41.2 will be treated presently. For details on the San Marco picture and its implications for a better appreciation of P41.2, see the following section. As it is, however, one must accept for now the juxtaposition of P41.1 and P41.2 on the same page by assuming that P41.2 constitutes a condensation or telescoping of the Calling, the Hiding, the Reproval, and the Denial of Blame, with the latter two components simply being taken for granted without explicit visualization. Under the interpretation given above that P41.2 explicitly depicts the Calling and the Hiding, the relative order of the two pictures on page 41 must be posited as follows: P41.2 (below) precedes P41.1 (top).120 This constitutes another instance of the bottom-totop sequentiality. While the bottom-up order is applied in other cases, too, this instance seems at odds with the preceding and succeeding pictures (P31, P34, P36, P39, P45): they all manifest the top-to-bottom order.

120 An alternative view is that the two pictures were misplaced: P41.1 (Judgment of the Snake) should have been placed below P41.2 (Raw 1953: 55). In addition to an unwarranted commitment to the absolute validity of the top-to-bottom order, this interpretation seems to create more problems than it solves. A fundamental problem in need of explanation is why the order was reversed by mistake in that particular place. The alleged misplacement would seem all the more questionable in view of the consistent topto-bottom arrangements on the adjacent pages, as treated in the text. By contrast, the postulation of the bottom-to-top order as in the current manuscript, as proposed here, is subject to principled explanations in overall and specific terms. https://doi.org/10.1515/9783110788068-010

218 

 10 The Judgment of Adam and Eve: P41/P44/P45.1

A question arises: Why was the bottom-to-top sequence adopted in defiance of the contrary order consistently implemented in adjacent pictures? On close inspection, it proves to be the case that, where two scenes of God’s acts are arranged along a vertical axis, they proceed from bottom up, as substantiated in P7 (section 4.1) and P9 (section 5.1), without counterexamples. Thus, the particular arrangement in P41 turns out to fully conform to the pattern detected. By contrast, the adjacent scenes all concern Adam and Eve on their own, or in relation to Satan’s emissary. Either way, God is not featured at all, with the exception of P45.1. In P45.2, the immediately following picture paired with P45.1, however, God does not make an appearance. Thus, inasmuch as God’s acts are not the focus of the page, the top-to-bottom sequence in P45 does not count as a violation of the generalization at issue. By way of statistical analysis, performing a Fisher’s exact test on the contingency table of two bottom up versus zero top down (for P7 and P9) and zero bottom up versus five top down (for P31, P34, P36, P39, and P45) gives a p-value of less than 0.05. It may follow then that the direction of vertical arrangement is significantly correlated to whether God figures as a central character on a given page. Another question to be answered is: Why did the artist combine P41.1 and P41.2 into one full-page illustration? Why did he not choose the alternative options of allocating the pictures to two separate pages, or of integrating P41.1 with P44 to form a full-page illustration? The latter option would otherwise have been optimal in that it would have made it possible to treat the thematically common tripartite scenes—God’s Judgment of the Snake, Eve, and Adam in that order (Genesis A 903a–938b)—in a single integral format. In this connection, it would be of paramount interest to point out that all seven pages running from P31 through P45, including P41 under examination, are composed of two pictures. The two alternatives mentioned above, however, would contravene this consistent pattern of allocation. On the one hand, to assign separate pages to P41.1 and P41.2 would create two single-panel illustrations, which would make exceptions to the generalization. On the other hand, to incorporate P41.1 into P44—the latter itself being bipartite already; see immediately below—would maximize the variation in picture constituency, ranging from one constituent (P41.2) to two (P31, P34, P36, P39, P45) to three (P44). The double figure of God in P44 is of vital significance here. Rather than placing a single figure of God at the center, the representation actually executed may highlight the duality in representation, whereby God’s judgments of Adam and Eve constitute separate events, as testified in the text. Thus, although not marked out by manifest means of demarcation, P44 comprises two distinct pictures. A single common figure of God, on the contrary, would inappropriately prompt a misinterpretation of God judging the couple in unison in defiance of the text and, more seriously, in violation of the allotment of two scenes to a single picture frame otherwise consistently implemented from P31 through P46, the series of scenes culminating in the Expulsion. As discussed in the preceding section, P41.2 is parallel to and contrasted with P31.2, first conceptually and then as a consequence, compositionally as well. While P31.2

10.1 A synchronic–structural perspective 

 219

depicts Adam and Eve calling to God in prayer and despair, P41.2 shows God addressing the couple, who are hiding under the trees. Thus inseparably interlocked and mutually dependent in their conceptual interrelationship, the two events constitute a pair of polar opposites, call and response. This underlying conceptual polarity would likely have induced a corresponding correlative visualization. Specifically, the establishment of the beginning moment—Adam and Eve Praying—would have inevitably activated the ending one—God Calling Back. P41.2 was thus composed in response to P31.2. By contrast, where the initial condition fails to materialize, the final consequence is unlikely to occur. It should be little wonder then that a comparable image of God Addressing Adam and Eve in response to their prayer failed to be produced in other places,121 in which analogous structural motivations were lacking in the absence of the scene of Adam and Eve Praying to God. As discussed in section 8.2 above, the Praying scene is demonstrably an innovation in Junius, unknown elsewhere in the Cotton family and/or the Touronian Bibles. In light of the structural opposition involved, it is only natural that P41.2 is substantially based on P31.2 in compositional terms. Of the three poses that the couple may take in the extended group of variant pictures spanning from P31.2 to P39.2, it is the unusual kneeling posture that is realized in P41.2 rather than the prevalent one of standing. While far less likely in purely statistical terms, the actual selection is firmly motivated: it is evidently based on the corresponding figures in P31.2. It should be noted that the Hiding is not necessarily correlated to kneeling, as exemplified at San Marco (Pl. 48). Of still further interest are the varying ways of kneeling between Adam and Eve. Eve is kneeling more deeply than Adam in P41.2. This difference may again be attributed to a comparable distinction that obtains in P31.2 and in an exaggerated form at that: Eve almost lies on the ground, her face down. Notable differences from the model picture of P31.2 are readily subject to explanatory accounts by reference to the artist’s presumable intentions and a set of compositional conventions and interpictorial dependencies, more specifically, the compatibility and harmony with preceding pictures. Since God is usually placed on the right side of the page in Junius, the orientation of Adam and Eve is reversed in P41.2 from P31.2, in which the privileged place is occupied by Satan’s messenger and accordingly the praying couple must face left, away from the devil standing at the extreme right. Moreover, unlike God’s opponents, Satan’s messenger in particular, God is usually positioned in an elevated place, thus accounting for the conspicuously higher position of God in P41.2 than that of Satan’s emissary in P31.2. Adam and Eve stand remarkably farther apart in P41.2 than in P31.2. This relatively long distance between the couple may be understood as a direct continuation of the immediately preceding picture (P39.2), in which the

121 As remarked by Broderick as follows: “. . . there is nothing analogous to what we see on page 41 of the Junius manuscript in any of the extant Carolingian Bibles, where the calling of Adam and Eve is uniformly eliminated in favor of the Accusation” (Broderick 1978: 229).

220 

 10 The Judgment of Adam and Eve: P41/P44/P45.1

pair are seated a comparable distance from each other. The trees are depicted in larger numbers and with richer branches and leaves in P41.2 than in P31.2. The luxuriant trees, in all likelihood, would have been motivated by the artist’s intention to express effectively the hiding act involved, as observed at the beginning of this section. Taking a more specific look at the trees, one is struck by a cluster of similarities between P41.2 and P39.2—just mentioned in regard to the distance between Adam and Eve—in their size and shape, and also in the ways their branches are extending above Adam and Eve. In the face of such a clustering of specific resemblances, it would be hard to deny that P41.2 would have been composed under the strong influence of the immediately preceding picture because of the pictorial contiguity and contextuality that the artist would have intentionally exploited for an effective representation based on a smooth transition from one scene to the other. In this way, essentially grounded on P31.2, P41.2 would have been composed through a host of rearrangements, largely determined by considerations both internal to the picture being created and external in the way of intervisual generalizations and contextual details. At the same time, the demonstrable continuity with P39.2, evidenced by a wealth of agreements in compositional details between P41.2 and P39.2, may indicate that P41.2, while primarily paired with P31.2, is still associated intimately with the group of the Covering (P34/P36/P39). Inasmuch as these pictures, including P31.2, are all traced back to the common scene type of the Covering/Hiding, the persistent synchronic linkage between P39.2 and P41.2 may seem hardly surprising, as it may be viewed as a corollary of their genealogical affinity through the intermediary of P31.2, in addition to their spatial contiguity. P41.1 and P44 depict the three separate but succeeding scenes of God judging the snake, Eve, and Adam one after another. The particular pose of God in P44—viewed in profile with the characteristic hand-raising gesture—is obviously based on P41.1, which, as an initiator of the series, thus sets the stage for subsequent scenes. Since the previous scene in which Adam and Eve appear is P41.2, a certain degree of compositional continuity is maintained in P44: God is standing on a hill below which Adam and Eve are present, looking up to God; Adam is raising one hand and covering himself with leaves held in the other; Eve is covering herself with leaves held in both hands. Since the couple is no longer hiding in P44, in contrast to P41.2, however, the trees under which they kept themselves concealed are removed from the scene. Also absent in P44 is an architectural frame despite its outstanding presence with an elaborate decoration in P41. Since P44 in all likelihood would have been intended as a continuation of and complementary half of P41 in general and P41.1 in particular, the artist would not have bothered to repeat details that are irrelevant to the major concern of P44, which focuses on the judgment of Adam and Eve to the exclusion of other details already depicted in the preceding illustration. Among those nonessentials are the trees on the margin—not only lacking the functional motivation as noted above but also obscuring the overall depiction itself—and the architectural frame—which is again irrelevant and can readily be filled in by referring to the preceding page. For the compositional reasons elucidated

10.1 A synchronic–structural perspective 

 221

above, the Judgments of Adam and Eve were conflated into a single format. In designing the merged version, the artist was confronted with a number of issues to address and a range of solutions to weigh against each other and choose from, which he worked out sensibly and competently, as shown in the following. According to the narrative order (Genesis A 903a–938b), P44 must be regarded as embodying a left-to-right sequence. These temporal dynamics in P44 account for the apparent exceptionality of Adam being positioned at the right on this page, contrary to the prevailing arrangement of Adam at the left in other pictures—notably P41.2— insofar as they are all static in the sense that no picture-internal successivity is demonstrably in force. Only a few apparent exceptions are observed in relation to this generalization. P9 is analogous to P44 by way of temporal development, although the sequential order involved is right to left. P10 and P45 follow a specific constraint to the effect that Adam stands closer to God than Eve. What would appear to be the only genuine exception to the generalization may be P28, in which Adam is placed at the right with Satan’s agent at the center distributing the fruit to Adam and Eve apparently at the same time. As substantiated at length in section 8.1 above, however, this picture is dynamic in representation: it depicts two successive events, first in the right half and second the left, with the central figure involved in both. Thus, P28 is parallel to P9: Adam stands at the right precisely because he is a participant in the earlier event, resulting in an apparent violation of the generalization that Adam is placed at the left unless otherwise motivated. At this point, a question arises: Why was the left-to-right directionality adopted rather than the converse, which is also independently applied, as in P9? Moreover, the reverse direction that was not chosen would have maintained the same positioning of Adam at the left and Eve at the right as in P41.2, whose compositional properties are, unless motivated otherwise, carried over to the following scene, P44, as observed above. The issue of directionality is not only of intrinsic importance on its own, but is bound to prompt an independent but closely related question of why a mirror-imaged arrangement (reflection) of Adam and Eve was executed, at the expense of a unidirectional concatenation (procession) such as is realized in P41.2. The above two issues must remain unexplained insofar as one sticks to P41.1 and P41.2 as the exclusive source material for P44. For their explanatory accounts, it seems indispensable to invoke an analogy with P28. In concrete terms, the composition of P44 would have been largely modeled after this picture in Junius 11, farther distanced from P44 than the latter’s immediate antecedents. A conceptual basis for the analogy thus posited may seem to be defensible with sufficient plausibility: while P44 represents the Judgment, that is, giving and receiving sentences, P28 depicts the Temptation, giving and receiving the forbidden fruit. The latter event led to the former through a chain of cause and consequence. Given such a conceptually motivated association, it should be only natural that a corresponding compositional link would have been forged on that cognitive basis. Drawing on the resultant visual linkage, one may answer the two questions raised in the last paragraph by attributing the relevant features to the model picture. In

222 

 10 The Judgment of Adam and Eve: P41/P44/P45.1

rough terms, P28 is vertically symmetrical or mirror-imaged, as far as the geometrical placement of Adam and Eve—ignoring their distinct identities, of course—is concerned, and Adam is placed at the right.122 The close connection between P44 and P28 claimed above may be corroborated by a partial similarity between P41.1 and P28. Specifically at issue are the double arches overhead commonly depicted in the two pictures. Since P41.1 and P44 are thematically in unity, and their separate presentation is no other than a matter of execution conditioned by contextual considerations, as discussed above, the resemblance at issue of P41.1 and P28 may be viewed as visual confirmation that the artist actually associated P44 with P28 in conceptual and compositional terms, as it is paired with and complemented by P41.1. Finally, the figuration of God in P44 needs to be addressed in rigorous terms. There are a number of options for depicting God. Seemingly the most faithful reproduction of P41.1, which sets out the model being claimed here, would be a single figure of God in profile at the center. Whether looking down at Adam or Eve, however, the resultant composition would be less than optimal as a depiction of God giving sentences to Adam and Eve separately, as only a single act of judgment can lend itself to an explicit representation. Speaking of a single depiction, a frontal view of God at the center would be most efficient, as partly demonstrated in P41.2, as it may stand evenly distanced and without favoring one person over the other. This option, however, must be ruled out flatly as contradictory to the fundamental decision that the artist would presumably have made earlier: the judgment scenes should be visualized in essentially identical fashion, modeled after P41.1, in which God is depicted in profile and shown looking straight at, and pointing his hand to, the person being sentenced. Put another way, the otherwise expected inheritance of pictorial features from P41.2 would have been blocked by the overriding force of the image of God judging established in P41.1, the immediately preceding scene that is incorporated into a visual triad with P44. Adhering to the succession of essentially the same three events, then, necessarily requires double figures of God addressing Adam and Eve, respectively. Although the double representation thus proves to be fully motivated and justified, details still need to be specified. At stake in particular is the configuration of the two figures of God. While each figure may broadly be defined as a reproduction of the figure of God in P41.1, their positioning relative to each other has yet to be determined. Concretely, should the two figures be combined together (and to what extent) or placed at a distance from each other (and how far)? Naturally enough, arranging the two figures wide apart would be hardly acceptable, as such a placement would detract from the ideal of unity and symmetry. Therefore, a proximate arrangement of the two figures presents itself as a viable option that is open to a spectrum of particular executions in details of configuration. At

122 Ultimately, one may have to appeal to the Junius artist’s pronounced predilection for symmetry—as permeated in the body of pictures examined so far.

10.1 A synchronic–structural perspective 

 223

one end of the spectrum, a complete separation of the two figures may militate against the unity of agency and the integrity of multiple acts. At the other extreme, a full amalgamation of the dual figures—particularly a merger of the heads—would have evoked objectionable associations with Satan’s messenger in P28, who, standing at the center and holding the fruit in both hands, tempts Adam and Eve in turn. It may be recalled that the central nonterrestrial being raising hands forward and backward is depicted as God’s opponent, as exemplified in P3.1, P28, and P31.2. Therefore, P28, already drawn in as a model to be partially reproduced, serves in the present case as an example of an inappropriate model to circumvent. At any rate, whether encouraging or discouraging, P28 would have borne heavily on the composition of P44, along with its immediate antecedents, P41.1 and P41.2. The particular double figure of God composed for P44, therefore, would have been produced by the artist’s careful and rational assessment of the precedents in P28, P41.1, and P41.2. In P45.1, God, standing at the upper right, and Adam and Eve on the opposite side at the upper left, fully dressed and facing God, are shown engaged in verbal interaction, all with pointing hand gestures. Behind and above the three characters is an architectural frame with an angled arch. Given the couple’s departure represented in the picture provided immediately below (P45.2; section 11.1), it may be reasonable to infer that the conversation concerns this event. Specifically, according to Gollancz (1927: xlii) and Ohlgren (1992: 93), P45.1 features God pronouncing the banishment of Adam and Eve from paradise. Directly continuing from the triple scenes of the Judgment in P41.1 and P44, P45.1 puts an end to God’s sentencing acts, thereby leading the way to the Expulsion (P45.2). This underlying conceptual link naturally groups these three pages closely, and may in turn justify treating them as a unit at a higher level of pictorial organization. A similar grouping has previously been identified involving P10/P11/P13 (Adam and Eve in paradise). In either case, the group members represent conceptually and temporally clustering states and events viewed from slightly different perspectives and/or with interests in distinct phases. The varying perspectives and concerns involved are naturally materialized by way of visual variations, which may be organized and systematized in one way or another. In this respect, comparing the two groups in question will reveal details of structural organization and material execution in ways that would be hardly accessible when an individual group is considered in isolation. It may be recalled at the start that the spatial order P10/P11/P13 correlates to a temporal sequence of events: God welcomes Adam and Eve in paradise; God beholds from above his creations including the couple; and finally Adam and Eve and other creatures are left on their own living on earth, while God is away and back in heaven. These developing phases are set in two differing frames of composition, which may be defined in regard to the configuration of Adam and Eve: arranged asymmetrically (P10) versus symmetrically (P11 and P13). To this parametric opposition is largely correlated that of God’s face orientation: profile (P10) versus frontal (P11). With P13—Adam and Eve symmetrically placed, but God viewed in profile—not fitting the pattern perfectly in this

224 

 10 The Judgment of Adam and Eve: P41/P44/P45.1

respect, P10 and P11 may adequately be identified as prototypical realizations of the two polar opposites, whereas P13 is conceptualized as a less than prototypical instantiation. With the above set of structural relations as a reference scheme, one may explore how the group of P41.2/P44/P45.1 is organized in its formal underpinnings and its realizations on the surface. By the primary defining criterion concerning the arrangement of Adam and Eve (asymmetrical vs. symmetrical), P45.1 and P44 correspond to P10 and P11, respectively. By contrast, the supplementary feature of God’s orientation (frontal vs. profile) fails to make the pair P44/P45.1 a perfect match with P11/P10, because of the independently required avoidance of a frontal face of God in P44. The remaining members, one for each group, namely P41.2 and P13, are parallel in that the two distinguishing features are in conflict for categorization of each: the asymmetrical configuration of Adam and Eve contradicts the frontal view of God in P41.2 on the one hand, and the symmetrical placement of the couple conflicts with the profile view of God in P13 on the other. As far as specifications of the two parameters are concerned, however, the actual ways of conflict are reverse between the two pictures. Due recognition of the structural correspondence between the two distinct groups, as substantiated above, may lead to a better appreciation of the details. Of particular interest is the presence and absence of an architectural frame and its specific structure. The correspondence pair of P45.1 and P10 is commonly provided with a large arch over the three figures, although the exact shape of the arch is different, angled (P45.1) or circular (P10). Standing at the other extreme, the set of P44 and P11 is lacking an arch structure altogether. Finally, the hybrid pair of P41.2 and P13 is topped with double arches. In both, in the central zone in which the two arches are joined, God is depicted in profile,123 and the figures of God—in the directly paired pictures, P41.2 and P13—do not act on Adam and Eve represented below. Still finer details concern Eve’s right hand in P45.1 and its compositional analog, P10: in both, Eve’s hand is depicted overlapping the left column.124 The intergroup relationship underlying these manifold compositional correspondences may well be identified as significant part of the pictorial organization that the artist would have drawn on in creating illustrations. In other words, one may postulate a cognitive reality behind the purely formal relationship. This cognitive interpretation would in turn bring about explanatory accounts of these agreements rather than dismissing them as simple accidents. Specifically, the artist would have built heavily on the group of P10/P11/P13, which he already must have brought into being, in composing the set of P41.2/P44/P45.1. In so doing, he would have taken the liberty of reusing similar materials already put to use for the model pictures. Such adaptations may have been simply due to artistic parsimony or economy. Or they may have been at least partly 123 Notice that the arches are directly over P41.1, in which God is shown in profile. 124 On the basis of this pictorial parallelism, Hilmo (2004: 91) recognizes an intervisual link between P10 and P45 and explores interpretation in terms of visual exegesis and metatextual reference. For a criticism of Hilmo’s exegesis, see Broderick (2009: 395–396).

10.1 A synchronic–structural perspective 

 225

intended as manifest expressions of the structural relations that the artist must have been fully aware of. Given that P44 and P45.1 are a symmetrical and an asymmetrical version standing in contrast, analogous to the variation between P11 and P10, they should correspondingly be identified as parallel to Pii and P2, respectively, because of the structural parallelism established between P10/P11 and P2/Pii in section 6.1 above. At stake here, however, is not to confirm the formal compositional relationship between the two groups, which should be self-evident by now and hence in no need of independent substantiation here. Of vital concern rather is to see how the recognition of such parallel relations makes principled explanations possible for a clustering of striking coincidences in specifics of composition that otherwise would be beyond understanding. The two faces of God looking sideways in P44 resemble those of seraphim in Pii. Of further interest, the left faces in both pictures are in three-quarter view, whereas the right faces are represented precisely in profile. Viewed from a wider perspective, the united figure of God flanked by the seraphim looks compositionally analogous to the double figures of God in P44. Specifically, the two figures of God in P44 that are unified except for the top section of the body, particularly the head, which is kept detached, create an overall shape that is found to be analogous to the configuration of God and the flanking seraphim in Pii, as well as to the hexagonal frame enclosing God in P11. All these three illustrations, characterized as symmetrical variants, thus are fully in harmony in abstract geometrical terms in their core representation of God. From this, one may venture to infer that the particular way in which the double figures of God are unified in P44 would likely have been designed carefully so as to optimally visualize the intergroup parallelism with the two antecedents, Pii and P11. P45.1 proves to be similar to P2 in its overall organization and compositional details. Standing at the upper right corner of the panel, God, in three-quarter view, faces to the left and is pointing with his right hand to Adam and Eve, who have sinned and who are positioned to the slightly lower left. Quite analogously, in P2 God looks to Lucifer, who has sinned (or is about to sin), who is standing in a comparable location relative to God. Adam’s pose in P45.1, particularly the gesture of raising both hands in front of the chest, is almost identical to Lucifer’s in P2. The major picture (P45.1, as opposed to P45.2) occupies the upper half of the zone, as in P2. A slight difference, however, is that the lower right is allocated to a separate picture in P45 (i.e., P45.2) but is left totally blank in P2 without a further image on the same page. The pair of P44/P45.1 is comparable to that of Pii/P2—as well as to that of P11/P10, remarked above—with respect to the representation of an architectural frame. Only in one member of each pair does such a frame appear. Moreover, the arches are angled in shape—which is extraordinary in terms of frequency of occurrence—in both pictures. Finally, by way of recapitulation and on a note of reservation, it must be pointed out again that the group organization of P41/P44/P45 does not exactly replicate those of Pii/ P2 and of P11/P10 in every detail. Of paramount interest, the opposition between frontal (symmetrical) and profile (asymmetrical) does not hold as rigorously for P41/P44/P45 as

226 

 10 The Judgment of Adam and Eve: P41/P44/P45.1

it does for Pii/P2 and P11/P10. While P45.1 may be viewed as being in profile (asymmetrical), a symmetrical counterpart is not available in its typical form. In P41.2 God is frontal, but his position, shifted to the right, is far from symmetrical, as are the nonconfronting figures of Adam and Eve. In P44, Adam and Eve are arranged symmetrically; so are the double figures of God. Yet God does not face front but turns sideways. Thus, these pictures might be conceptualized as two nonprototypical instantiations of the frontal, symmetrical type. It should also be borne in mind that unlike Pii/P2 and P11/P10, not all member pictures are self-standing illustrations. This means that, even if instantiated in a prototypical manner, the embodiment of relevant properties at issue must be compromised because of an accompanying picture of a different nature, as with P45.2 below P45.1.

10.2 A diachronic–comparative perspective As Broderick (1978: 231) notes with reference to Green (1955: 345), the Judgment of the Serpent, Eve, or Adam was not visualized on its own in the Cotton Genesis tradition, in contrast to a preceding scene of the Reproval/Denial of Blame, which is extensively illustrated. The only exception to this generalization is San Marco (Pl. 50): in it, Adam and Eve, facing each other, are kneeling before the enthroned God, who, in three-quarter view, is sentencing the snake, which, head down, is about to leave the scene. The San Marco scene outlined above seems of enormous interest in exploring P41.1, P41.2, and P44 from a comparative and diachronic perspective. Impressionistically, it may be characterized as a condensation of the three Junius pictures in question into one: God’s judgment of the snake is based on P41.1, the kneeling figures of Adam and Eve derive from P41.2, and the symmetrical arrangement of the couple owes to P44. A better appreciation, however, must await detailed examination below. Since the Hiding is separately represented at San Marco (Pl. 48), and since there is no sign detectable for inferring a hiding act in the San Marco picture at issue (Pl. 50), the couple cannot be equated with the P41.2 counterpart in thematic terms. Rather, it should be identified with the P44 figures of Adam and Eve, who are receiving sentences. The San Marco picture, however, does not literally depict God’s judgment of the couple; it provides only a scene of the judgment of the snake, with the couple kneeling beside. The scene involving the couple at San Marco may be amenable to two alternative interpretations. First, although the judging God is not explicitly shown engaged in giving sentences to Adam and Eve—conceivably due to lack of space—they are each actually being sentenced, as seems quite plausible from their pose and facial expression. Such a disparity between conceptualization and visualization may be subject to a plausible account. To repeat the figure of God three times as logically required would necessarily not only have resulted in a rare occurrence of multiple figures of the same being in a single panel but also have incurred other undesirable consequences of having to substantially reduce the size of God and detracting from a symmetrical configuration of God, Adam, and Eve such as is realized in its current form. Thus, the single figure of God

10.2 A diachronic–comparative perspective 

 227

judging the snake anticipates his subsequent two acts of gazing at and pointing to Eve and Adam in turn. The picture is therefore a complex of three distinct scenes in which Adam, Eve, and the configuring God and snake belong to the three differing temporal phases of the Judgment. Alternatively, the kneeling Adam and Eve may be viewed as having yet to be sentenced; that is, they are waiting for their turn to be judged, while God is still acting on the snake. The couple’s submissive posture and humble look appear to be as compatible with this reading as with the previous interpretation. On this second reading, then, the picture is monolithic in temporal terms: it would have been intended to be a momentous representation of an event at a single point in time rather than a conflation of succeeding multiple scenes. Since San Marco is generally distinguished by a constellation of pictures each specifically allocated to single scenes without overlap or omission, to read into the picture under consideration a succession of distinct events or consecutive phases of an event would seem to be a grave flaw of not doing full justice to one of the compositional principles governing the San Marco mosaics. In this light, the second interpretation should be chosen as the more feasible one. Regardless of which interpretation may be adopted eventually, however, one can afford to engage in a formal inquiry of the San Marco picture in its own right. As suggested earlier, it may be characterized as a conflation of P41.1, P41.2, and P44 in purely formal and compositional terms. At stake here is whether a derivational relationship may be postulated between the two works, and, if so, which way the derivation proceeded, from Junius to San Marco, from San Marco to Junius, or both derived from a common source. Beginning with the snake, while in P41.1 it is depicted in two different forms sequenced along the time axis (first the upright snake receiving the sentence, and then the crawling one leaving the scene upon judgment), it is represented only once at San Marco—as expected—in the reversed head-down form, which may be viewed as a snapshot taken at a moment of the snake beating a hasty exit from the scene after being sentenced.125 In terms of temporal sequence and physical movement, the San Marco snake is thus placed in transition between the two figures in P41.1. In other words, it is a compromise that would have been reached by mediating the two phases of the snake’s action. Seen in its own right as an autonomous representation of God judging the snake, the San Marco depiction would be found less than optimal and natural, with the accused not confronting the accuser. To account for this lack of a perfect match between conceptual and visual in the San Marco case, one would assume that the state of affairs there would have been a forced, secondary result, that is, a consequence of working out an acceptable solution by negotiation and compromise in the face of conflicting requirements and conditions. The assumed reworking may have been caused, among others, by

125 In this connection, Jolly’s (1997: 55) following observation does not seem very illuminating: “at least its [the Serpent’s] lowered head demonstrates its inferiority.”

228 

 10 The Judgment of Adam and Eve: P41/P44/P45.1

one of the principles that constrain the composition of a panel at San Marco. As stated above, depicting more than one figure of the same being in a given frame of illustration is avoided at San Marco, if not strictly prohibited. There are only two exceptions to the rule: first, the scene of God Removing a Rib from Adam (left, Pl. 43) and God Forming Eve (right); second, the scene of Eve Picking the Fruit and Eve Tempting Adam (Pl. 46). The situation under which the San Marco illuminator had to work may be reasonably attributed to the challenge he faced of adapting a model like P41.1 to the compositional requirements imposed on his current work. In short, it would make sense to assume that the peculiar form of the snake would have emerged as a compromise between the two distinct figures depicted in the model. Whether P41.1 served as a direct model or not, however, is a separate issue requiring other considerations beyond the present concern here. By contrast, the contrary trajectory of an originally single figure being split into two such as are manifested in Junius would be contrived and complicated, particularly given that the supposedly initial point—the head-down and head-away posture—may well be rejected as conceptually unnatural in the absence of specific evidence to the contrary. Turning to the figure of God, three features deserve special attention: first, the profile/three-quarter view, as opposed to frontal; second, the leftward, rather than rightward, orientation; and third, seated, as opposed to standing. While the first feature of the profile view of God is readily derivable from P41.1 and P44 but definitely not from P41.2, the third point, the enthroned figure, is not attested at all in the triple representations in Junius. In San Marco, a similar seated posture makes its appearance only immediately before the scene in question, the Reproval of Adam and Eve (Pl. 49), after it occurs for the first time together with Adam in the scene of the Naming of the Animals (Pl. 42). Thus, the present issue may be transferred to the previous scenes for a fuller understanding, not least because of the more frequent occurrences of the corresponding scene in related works (Table 12). In any event, even seen from such a broadened perspective encompassing the Reproval, the enthroned God must be viewed as an innovation unique to San Marco, and as such it would lack pertinence to the present concern with exploring the relation between Junius and San Marco.126 It is then the second property of the leftward orientation of God in the Judgment scene at San Marco that is worth inquiring about with respect to its possible link between Junius and San Marco. From the creation of Eve (Pl. 43) through the Reproval/ Denial (Pl. 49) at San Marco, God is depicted consistently facing right, as in the scenes following the Judgment. In view of the exceptionality of God facing left, and by contrast the ubiquitous appearance of this orientation in Junius including P41.1, it may be legitimate to assume that the exceptional orientation at San Marco would have been introduced 126 Compare Kessler (2009–2010: 23): “The Creator, here enthroned to emphasize his role as ruler, is a variant of the standard type that recurs also in the Denial of Guilt and Curse of Adam and Eve . . .” See also Jolly (1997: 54): “God’s enthronement like a presiding judge, another detail altered from the fifth-century model, confirms the gravity of the moment and emphasizes the theme of accusation, trial, and judgment.”

10.2 A diachronic–comparative perspective 

 229

from a model like P41.1 in Junius. This supposition has its plausibility enhanced by the finding adduced above that the other constituent of the two major figures in P41.1, namely, the snake, played a significant part in shaping its correspondence at San Marco. Put together, the scene of God judging the snake would in essence have been a wholesale borrowing from a model that is faithfully preserved in P41.1. Adam and Eve receiving or about to receive sentences are shown kneeling at San Marco, in contrast to their standing posture in P44, a corresponding scene of the Judgment in Junius. Should this kneeling pose at Judgment be counted as another instance of compositional innovation that was introduced exclusively at San Marco? It is at this juncture that P41.2 may be brought to bear on the issue. In it, too, Adam and Eve are comparably kneeling, but they are hiding behind trees from God’s calling, not being judged or awaiting sentences to be given in the presence of God. Since the corresponding picture of the Hiding at San Marco (Pl. 48) exhibits the couple standing rather than kneeling, one may be led to infer that the kneeling may have been adopted from external sources. Thus, there is a possibility that through the conflation of complex consecutive scenes the kneeling pose would have made its way into the Judgment scene at San Marco, originated from the Hiding/Calling, such as is attested in Junius (P41.2). A simple transfer like this, however, would likely take place preferably in the presence of favorable conditions on the part of a recipient. It is accordingly appropriate to take a closer look at the situation at San Marco. In contrast to the Hiding scene at San Marco, Adam and Eve are depicted crouching in the scene of the Reproval/Denial (Pl. 49) that directly precedes the Judgment picture. The crouching pose is limited to this scene at San Marco, much as the kneeling one is unique to the Judgment. That these two unique poses are arranged in succession should be far from coincidental. How do these two poses relate, then? With reference to the varying degree of bending the knees, crouching and kneeling may be grouped together by virtue of the marked feature of bending the knees in opposition to the unmarked counterpart of standing with the knees straight. To restate in more accessible terms, crouching may be characterized roughly as a midpoint along a scale whose extremities are occupied by standing (no bending) and kneeling (maximal bending). In light of this overall scalar distinction, then, kneeling would readily lend itself to being conceptualized as a reinforced version or advanced stage of crouching. Given that kneeling and crouching thus differ from each other only as a matter of degree, one may conceive that the apparently unique innovation of representing the kneeling couple in the Judgment scene at San Marco may well have been a logical conclusion of bending the knees, the practice already activated in the immediately preceding picture. Thus, with San Marco being ready on its part to appropriate the kneeling couple, the contact with a model like P41.2 would have triggered actualization of the latent image there. The symmetrical arrangement of Adam and Eve is another outstanding feature. Up to this panel (the Judgment, Pl. 50), San Marco exhibits no comparably symmetrical, face-to-face placement of the couple. There would have been nothing wrong with depicting the kneeling couple facing the same direction, analogous to the Reproval scene

230 

 10 The Judgment of Adam and Eve: P41/P44/P45.1

(Pl. 49). In need of explanation, therefore, is the unique occurrence of the symmetrically configured couple in the Judgment scene of San Marco against the background of rare visualization of the same scene elsewhere irrespective of how the couple is arranged. In the absence of competing source material to choose from, P44 (or its cognates now lost) would suggest itself as the only viable model, insofar as one remains intent on identifying a source for explanation. In this connection, it should be kept in mind that, as noted above, San Marco exhibits no marked preferences for symmetry, in sharp contrast to Junius, in which such a predilection is pervasively in force. Therefore, an appeal to an internally motivated elaboration culminating in symmetry would be hardly as plausible for San Marco as for Junius. The positioning of Adam at the left and Eve at the right, which is contrary to the placement implemented in P44 but conforms to the arrangement in P41.2, may readily be ascribed to the same pattern that obtains in the preceding scenes from the Covering (Pl. 47) through the Reproval (Pl. 49) at San Marco, without going out on a limb to invoke P41.2 as a source of inspiration. The relevance of P41.2 as an explanans is not only dispensable but also dubious. As argued in the preceding section, the relative positioning of Adam (right) and Eve (left) in P44 would have been due to the compositional conditions that the Junius artist specifically followed. In the absence of analogous constraints imposed on the San Marco artist, he would have found no reason to reverse the prevalent pattern—Adam (left)/Eve (right)—in the Judgment. In summary, as far as no direct contact relationship is recognizable between Junius 11 and the San Marco mosaics (Broderick 2014: 223), the Judgment at San Marco provides invaluable testimony that P41.1, P41.2, and P44 in Junius 11 derive from the Cotton Genesis tradition, the same family that San Marco is affiliated with. Specifically, the particular configuration of God and the snake, the symmetrical arrangement of Adam and Eve face to face, and their kneeling pose—the major properties that are common to Junius 11 and San Marco—would unlikely have evolved independently and accidentally. Rather, they may legitimately be traced back to the common source in the pictorial tradition that both works instantiated in their own different ways. Furthermore, the ways in which these commonly attested features are condensed in a single complex representation at San Marco would suggest that the representation there would have developed through reorganization and elaboration at a later stage of evolution. In the absence of other corresponding examples in the Cotton family, the simplest possible assumption one may conceive would be that the Judgment scenes would have been introduced by innovation in Junius. Correspondingly and specifically, the Judgment at San Marco would have been derived from P41 through the superimposition of the upper and lower halves, as will be further defended in section 12.4 below.127

127 For a concomitant reconsideration of the membership of the Judgment series of pictures in Junius, see below after the discussion of P41.2. For implications of the Junius innovation of picturing the Judgment, see further below.

10.2 A diachronic–comparative perspective 

 231

However, by far more striking than the similarities between Junius and San Marco discussed above are the resemblances found between P41.2 and Bamberg 7v2 (right). Adam and Eve are covering their nakedness. Adam is situated at the left, and Eve at the right. The two face the same direction, rightward. Large branches hang over the couple. Adding P31.2 to the pair, the similarities among the trio become more salient, as brought to light through a comparison of P31.2 and Bamberg7v2 in section 8.2 above. Given the two-pronged correspondences of Bamberg 7v2 to P31.2 on the one hand and to P41.2 on the other, it seems conceivable that the Junius artist would have composed a pair of contrastive pictures—P31.2 and P41.2—from a model resembling Bamberg 7v2. The two Junius pictures may be differentiated by varying degrees of symmetricality. In brief, P41.2 exhibits a greater degree of symmetry. The three figures—God, Adam, and Eve—are arranged in a well-balanced triangular form. Although facing the same direction, Adam and Eve are similarly kneeling and looking up toward God, who is shown in frontal view, standing at the top and slightly off center. The three trees are placed evenly and enclose the couple in parallel fashion, with the largest one at the center and the smaller ones on the margins. In addition to the properties bearing on overall symmetry, the two Junius pictures, together with the Bamberg one, are contrasted with each other along a few parameters: the couple’s orientation, right (P41.2 and Bamberg) versus left (P31.2); God’s presence (P41.2) versus absence (P31.2 and Bamberg; devil’s presence is immaterial); degrees of bending the knees, shallow (crouching: P41.2 [Adam] and Bamberg) versus deep (kneeling: P31.2 and P41.2 [Eve]). In light of the closer correspondences to Bamberg 7v2, P41.2 might be regarded as an unmarked or more faithful reflex of the model. Yet the presence of God and the relative distance between Adam and Eve might argue against this characterization. The presence of the devil, however, makes P31.2 more deviant than P41.2 from the model. The absence of God in paradise in Bamberg 7v2 entails his presence in heaven above, but it absolutely does not imply the presence of Satan’s messenger by any stretch of imagination, at least in terms of inherent narrative contents according to the biblical tradition. Thus, the depiction of God in P41.2 adds nothing foreign and unimaginable to the scene explicitly represented or implicitly presupposed in Bamberg 7v2, whereas the appearance of the devil constitutes a radically different event. It should be noted that the couple is kneeling in P31.2, Eve alone is kneeling in P41.2, and both figures are crouching in Bamberg. The kneeling pose is independently attested in the Judgment scene of San Marco, as discussed above. The two cases of kneeling in P31.2 (Adam and Eve) and P41.2 (Eve), however, can be distinguished in terms of relative similarities to the crouching pose in Bamberg, as measured along the scale of varying degrees of the knees being bent, referred to above: P41.2, with a lesser degree of bending the knees, is counted as comparatively closer to Bamberg 7v2, as also confirmed by the crouching figure of Adam in P41.2 much as in Bamberg. Of related interest, Eve is differentiated from Adam by a greater degree of bending the knees, in parallel in P31.2 and P41.2. All in all, then, it still can be maintained that P41.2 reflects Bamberg 7v2 more transparently.

232 

 10 The Judgment of Adam and Eve: P41/P44/P45.1

The above attempt to differentiate P31.2 and P41.2 with respect to their varying similarities to the Bamberg model and determine their relative fidelity to the exemplar seems to be utterly misconceived, however. As may be recalled, P31.2, along with P34.2, would have been indebted to Bamberg 7v2 and its cognates for inspiration (section 8.2). The binary variation on the same motif—the Covering/Hiding—was split into two distinct scene types, the Praying (P31.2) and the Covering (P34.2). Thus redefined upon emergence as a depiction of Adam and Eve calling to God in prayer, P31.2 in turn would have given impetus to the creation of P41.2 as a conceptual and compositional complement, God addressing the couple in reply. It is far from true that both P31.2 and P41.2 were directly brought into being by Bamberg 7v2. Rather, P41.2 is indebted to P31.2 as an immediate source of inspiration. Therefore, what should be at issue is the way in which P41.2 came to be shaped in response to the stimulation of P31.2 as a matter of derivation and organization that are internal to the pictorial system of Junius. Now that the question has been reformulated, one may properly address the issue by reiterating the greater symmetry exhibited in P41.2 than P31.2. As it turns out, this is a familiar, recurrent contrast between symmetry/frontal and asymmetry/profile. By reference to this overarching principle of pictorial organization underlying the Junius system, it may be assumed self-evidently that the two pictures in question are paired on the basis of the symmetry versus asymmetry variation. That is, once a new picture (P31.2) has been created, it needs to be positioned firmly in the system by being incorporated into the network of association, conceptually and compositionally defined. Thus, P41.2 should be appreciated primarily as a system-based, creative composition, as applies extensively to the construction of other pictures (Table 15, section 12.1). What seems to be unique to P41.2, however, is the discontinuous pairing between P31.2 and P41.2, while other comparable associations involve a series of images distributed in sequential continuity. The derivational dependence of P41.2 on P31.2 thus established in turn may prompt one to reconsider the group of P41.1/P41.2/P44 so far assumed to be in unity rather naively due to the integral coupling of the former two (P41.1/P41.2) into a single picture (P41). As it turns out, however, P41.1 and P41.2 should be separated in such a categorial way that P41.1 and P44 are kept fully integrated to the exclusion of P41.2, which is paired with P31.2 conceptually and compositionally, and is secondarily associated with the Covering group (P34/P36/P39), particularly its most peripheral member, P39.2, not least motivated by their spatial proximity (i.e., P41.2 is immediately preceded by P39.2 in terms of visual narration). Thus having been delineated properly, the whole group of P41.1/P44 may require reconsideration and reevaluation in a new diachronic light. In the previous section, it was claimed that P41.1, the Judgment of the Snake, largely determines the pictorial framework of the triad Judgment scenes (P44, left and right) by serving as the initiator of the series. Then, as substantiated in the course of discussion in the present section, the Junius representation of the Judgment was identified as an innovation, as it is older than the other reflex of the same scene at San Marco in art historical terms. Corre-

 233

10.2 A diachronic–comparative perspective 

spondingly, the latter must be characterized as derivative of the former. Now a question arises: Given that the Judgment scene emerged for the first time in Junius, from where did it originate in genealogical terms and why and how did the innovation take place particularly in Junius on structural and cognitive grounds? While the second question was already addressed from a synchronic perspective in the preceding section, toward providing a principled account for the first issue, one must take into consideration the whole distribution pattern of the Reproval, the Denial of Blame, as well as the Judgment in the major cognate works, as provided in Table 12.128 Before moving forward, however, one may pause to be reminded that the three scenes at issue are determined on the following pictorial grounds: – – –

Reproval (Gen. 3.11): God raises his hand in rebuke to Adam and Eve. Denial (Gen. 3.12–13): Adam and Eve point to Eve and the snake, respectively. Judgment (Gen. 3.14–19): God punishes the snake, Eve, and Adam individually and distributively in this order.

Thus, while the Reproval requires Adam and Eve in pair, the Denial and the Judgment involve the trio of Adam, Eve, and the snake as addressees. Moreover, while God looks to the trio collectively in the Denial, he faces each individually in the Judgment, at least susceptible to such an interpretation in visual terms. Table 12. Distribution of the Covering/Hiding, the Reproval, the Denial of Blame, and the Judgment among the major members of the Cotton Genesis family Junius

Episode

Praying P31.2

Covering/Hiding

Covering P34.2

P41.1/P44

B 

Reproval

√

SP

H 

√

√

√

√

Denial of Blame Judgment

V 

SM

M 

√ Covering

√

Calling /Hiding P41.2  

G 

 

 

√ Hiding

 

 

 

 

√ 

HD

 

√

√   

SPF

√

√   

√ 

 

 

 

 

Key (for details, see section 1.2): G = Grandval; B = Bamberg; V = Vivian; SP = San Paolo; H = Hildesheim; SM = San Marco; M = Millstatt; SPF = San Paolo fuori le Mura frescoes; HD = Hortus Deliciarum.

128 For exploring empirical implications of the proposed correspondences among the Cotton Genesis family and formulating a principled account of the Junius innovation of picturing the Judgment scene on its own, see section 12.4 below.

234 

 10 The Judgment of Adam and Eve: P41/P44/P45.1

As should be evident in Table 12, in none of the Cotton family members are the Reproval, the Denial of Blame, and the Judgment differentiated fully in a three-way contrast. At most two of the three episodes are distinguished: San Marco makes a distinction between the Reproval/Denial (Pl. 49) and the Judgment (Pl. 50), while the San Paolo frescoes separate the Reproval (fol. 27r; https://digi.vatlib.it/view/MSS_Barb.lat.4406/0028; Kessler 1977: fig. 27) and the Denial (fol. 28r; https://digi.vatlib.it/view/MSS_Barb.lat.4406/0029; Kessler 1977: fig. 28). In the remainder, a single scene encompasses two episodes by conflation and abbreviation, whereby the distinction between the Reproval and the Denial of Blame is neutralized and unified into one and the same picture, as exemplified in Bamberg and also San Marco. Alternatively, either the Reproval or the Reproval/Denial of Blame is integrated with earlier episodes, notably the Covering/Hiding. On the one hand, in Vivian (Pl. 27), Millstatt (fol. 12r, Pl. 65), and the Hortus Deliciarum (fol. 17v, Pl.  69), only the Reproval is combined with the Covering/Hiding, to the exclusion of the Denial, which is categorically absent. On the other hand, in Grandval (Pl. 24) and Hildesheim, the conflation extends from the Covering to the Denial. By far the most extensive in conflation is the San Paolo Bible (Pl. 34), which subsumes all episodes from the Temptation of Eve through the Reproval under a single scene. This comparative perspective brings to the fore the peculiarity of Junius in its treatment of the Reproval and the Denial of Blame. These two episodes fail to receive visualization in Junius, in contrast to the remainder, in which at least one of them is represented, either on its own or in combination with other episodes. Therefore, one is required to provide a principled explanation for this singular gap (as symbolized by the void second cell from the bottom in the leftmost column, Table 12), a unique feature of Junius unknown to other cognate works. At the same time, it must be recalled that, as substantiated above, Junius would presumably have introduced a depiction of the Judgment scene, an innovation unprecedented in the Cotton Genesis family, which was subsequently adopted at San Marco. Thus, one is confronted with a complementarity of two features unique to Junius—the absence of the Reproval/Denial of Blame on the one hand, and the emergence of the Judgment on the other. The clustering of these two complementary properties, accordingly, may lead one to the following most reasonable inference: the two phenomena are inseparable from each other, constituting two sides of the same coin. Specifically, the emergence of the Judgment follows from the absence of the Reproval/Denial of Blame. In other words, the novel scene would have arisen in Junius through reinterpretation of the inherited representation of the Reproval/Denial of Blame as one of the Judgment. Probable material for this recomposition can be specified in more rigorous terms. Given that a reference to the snake is indispensable in the initial act of the Judgment, as featured in P41.1, the involvement of the component of the Denial of Blame seems essential, as the Reproval proper concerns Adam and Eve alone (as exemplified in Vivian, Millstatt, and the San Paolo fuori le Mura frescoes, where no figure of a snake makes an appearance). Accordingly, the most plausible model for the Junius innovation would have been Bamberg, Grandval, and Hildesheim, in which not only Adam and Eve but

10.2 A diachronic–comparative perspective 

 235

also the snake are featured. Among the three candidates, Bamberg would appear closest to Junius in structural terms, not because the figures depicted are maximally similar, but mainly because it distinguishes between the Reproval/Denial and the Covering, as does Junius. As for the representation of the snake itself, the pivotal figure of the Judgment series initiated in P41.1, its outstanding double visualization in Junius would presumably have owed to both Bamberg and Grandval, with the standing figure derived from Grandval and the creeping one from Bamberg.129 As far as the figuration of Adam and Eve in P44 is concerned, Junius and Grandval seem to be matched most closely: in both Junius and Grandval, they are shown standing straight, whereas in Bamberg and Hildesheim the couple stoops markedly. Moreover, the couple’s apparently weird posture in Junius becomes understandable in reference to the Grandval underpinnings. That is, while the couple’s faces are viewed in profile in Junius, their trunks are oriented forward to the viewer and their legs turned leftward. Significantly enough, Grandval exhibits well-matched body orientations, notably with Adam. While the figures of Adam in Junius and Grandval are largely analogous, more remarkable is a conspicuous resemblance between Junius’s Eve and Grandval’s Adam with respect to their body orientation excluding the head and the hands, which are horizontally reversed. Accordingly, it may be conjectured that the left half of P44 involving God and Eve was derived from the configuration of God and Adam in Grandval by horizontal reversal, much as P41.1 constituted a mirror image of God and the snake located in the same Grandval picture. All considered, then, the Junius artist would have adapted the Reproval/Denial of Blame scenes in Bamberg and Grandval—partially in mirror image—as a representation of the Judgment, and made necessary adjustments in accordance with the new signification inscribed and in harmony with the syntagmatic relationship with the preceding pictures—particularly P28 and P41.2—in the same manuscript. The motivations for the pictorial conversion that inspired the Junius artist, and the whole operation of reorganization, reconfiguration, and rearrangement that he carefully executed toward his objective were already discussed from a synchronic–structural perspective in the previous section.

129 At Hildesheim one finds what appears to be a quadruped instead of a snake.

11 The Expulsion: P45/P46 11.1 A synchronic–structural perspective The Expulsion scene is visualized on two pages, P45 (Pl. 21) and P46 (Pl. 22), and each page consists of two parts. P45.1 depicts God making a final pronouncement on Adam and Eve’s expulsion from paradise, as treated in the previous section. Thereupon, the couple leaves through the gate of paradise, as represented in P45.2, in the lower right of page 45. In both parts, Adam and Eve are fully dressed, implying, along with their naked depictions up to P44, that they have just been clothed, although a corresponding scene is not pictured in Junius (compare with San Marco, in which the dressing scene is provided immediately before their expulsion: The Clothing of Adam and Eve, Dumbarton Oaks Research Library and Collection SM0798, https://iiif.lib.harvard.edu/manifests/ view/ids:12922122; Demus 1984: pl. 129). In departure at the gate, Adam holds a spade in his right hand and a bag of seed in his left, while Eve has what may be identified as a spindle whorl in her right hand (Gollancz 1927: xlii; Ohlgren 1992: 93; Broderick 1978: 233).130 At the left side of the following page (P46.1), an angel, standing just behind the doorway and brandishing a sword, guards the gate of paradise. At the right (P46.2), having left paradise for the outer world, Adam and Eve are standing in the open field; as in the earlier scene (P45.2), Adam holds the two implements for farming but Eve does not seem to carry anything. The scene of God sentencing the couple to leave paradise, as depicted in P45.1, is unprecedented elsewhere in the Touronian Bibles and in the Cotton Genesis family, although God pushing out the couple is occasionally represented (for details, see the following section). This innovation, unique to Junius, of depicting God making a final pronouncement may be ascribed to the generative power of the variation between static/ frontal and dynamic/profile that permeates the pictorial organization of the Old English/ Old Saxon Genesis under investigation. Specifically, P45.1 may be conceptualized as an asymmetrical/dynamic variant in opposition to P44—symmetrical/static—in parallel to other paired opposites, especially the contrast between P10 and P11. Now that P44 has been brought into being, an asymmetrical counterpart needs to be created as its complement. Seen from a different angle, P10 and P45.1 are characterized as polar opposites, entry into paradise on the one hand, and expulsion from paradise on the other. In this light, this conceptual opposition also would have motivated a composition that was materialized as P45.1—God’s final call for expulsion from paradise—by analogy to P10, God’s initial call for entry into paradise. In either case, the absence of a picture like P45.1

130 For a further association with the fruit on the one hand and the typological relationship to Mary and Sarah on the other, see Karkov (2001: 76–77). Of particular interest is the parallelism to the figure of Sarah carrying a whorl in P88, Abraham Approaching Egypt (Gollancz 1927: xlvi; Karkov 2001: 77). https://doi.org/10.1515/9783110788068-011

238 

 11 The Expulsion: P45/P46

would have been determined as wanting on structural grounds and hence objectionable in terms of the overall pictorial organization that the artist was exploring. In contrast to San Marco (Pl. 51), Millstatt (Pl. 66), and the Hortus Deliciarum (Pl. 69, lower right), the scene of God pushing the couple out of the gate does not receive explicit representation in P45 but the Junius picture is compatible with the San Marco/Millstatt interpretation: along with the couple, God can be conceived of as relocating toward the gate and driving them out.131 In this connection, it may be recalled that the central figure of the devil in P28 plays a double role, one succeeding after the other, first as Adam’s tempter and then as Eve’s counterpart. In parallel to this but admittedly in less transparent form, the figure of God in P45 may be construed as secondarily engaging with Adam and Eve’s departure from paradise. Of greater importance, however, is, whether physically pushing or not, God has dictated the couple’s expulsion. By depicting the consequent event of the couple’s departure in the lower zone of the same page, P45 asserts without qualification God’s absolute agency in expelling Adam and Eve. This assertive force is motivated in structural terms within the pictorial system that the artist is working on. At stake is a structural parallelism in visualization. The causal relationship between P45.1 and P45.2 is analogous to that between P3.2b and P3.3, and between the upper and lower halves of P16: the angels or humans disobeying God are driven out of their original dwelling place (heaven or paradise). On this basis, there would have been no need to depict explicitly God expelling the couple in P45.2. Without a manifest representation, God’s agency behind the event is fully appreciable. As disloyalty to God entails vital consequences for the antagonists, a prior confrontation scene suffices to convey the needed message. It should be noted that the recognition of the pictorial parallelism hinges on the subsumption of P45.1 under a conceptual category common to P3.2b and P16.1, the asymmetrical/profile variants with dynamic agency. The formal–conceptual parallelism to P3 and P16 is not confined to the relation internal to P45. A similar linkage also extends to cover a larger domain involving P45/ P46 on the one hand and P3/P16/P17 on the other. The fallen Lucifer is stuck at the gate of hell in P3; in subsequent pictures (P16.2/P17.2), he is restrained at the bottom of hell. Similarly, P45.2 depicts Adam and Eve on the point of departure from the gate of paradise. P46 then shows them out in the field at some distance from paradise. The double depiction of expulsion is thus motivated on internal grounds for securing pictorial parallelism between analogous scenes so as to optimize the whole organization. The two respective representations involved are concerned with two sequenced phases of the same event. Of further interest regarding the interface of P45 and P46 is the relative positioning of Adam and Eve, which is reversed in P46: in P45.2, Adam stands at the right as natu-

131 The movement of Adam and Eve from P45.1 to P45.2 is analogous to that of the snake in P41.1. Upon receiving God’s pronouncement, the accused take leave.

11.1 A synchronic–structural perspective 

 239

rally expected given his corresponding position in P45.1, whereas in P46 Adam occupies the left position relative to Eve contrary to expectation based on pictorial continuity. One may wonder whether this alteration was intentional or simply accidental. This change in positioning seems to have been internally motivated. As suggested, the positioning in P45.2 is an inheritance from P45.1, in which Adam stands closer to God than does Eve, as usually arranged. The reversed position in P46 in turn may be ascribed to the resultant shorter distance between Adam and the angel, who may be characterized as comparable to God because of his role as God’s proxy.132 As detailed below, the backto-back orientation of Adam and Eve in P46.2 is motivated through opposition to P13, in which the couple stands face to face. Notice that the spade and the bag are reversed (Broderick 1978: 242); the left hand holds the bag in P45 but the right one does so in P46. This change may have to do with the pointing gesture involved. Moving onto the composition of P46 proper, P46.1 strongly evokes P9 (Broderick 1978: 237; Karkov 2001: 56, 79). Particularly similar, in addition to the figure of the angel itself, are the form and decoration of the gate and the door. Against the background of these common features, notable contrasts stand out. While P9 depicts Michael frontal, P46 represents the angel in three-quarter view. This contrast between frontal and nonfrontal logically correlates to that of symmetrical (P9) and asymmetrical (P46) with respect to the representation of the angel. The contrast in symmetricality, however, also applies to the entire page in terms of pictorial composition. In P9, Michael stands just beyond the ladder and watches the creation scene in paradise, which unfolds on both sides that are separated by the central axis formed by the ladder (section 5.1). The overall composition is accordingly symmetrical. By contrast, in P46 the angel turns rightward to Adam and Eve. Furthermore, while the left side depicts paradise, the right one represents the world outside paradise. All these heterogeneities contribute to an asymmetrical composition of the whole image. Of more substantive significance is the opposition between static (P9) and dynamic (P46). The greater dynamicity inherent in P46 is evinced by the upraised sword held by the angel as well as its pose, particularly the posture of its lower body. The recurrent clustering of these familiar features may justify characterization of these two pictures as a pair of variants in line with the well-established group exemplified by the representative set of Pii/P2. While all the members identified so far concern God as subject of action/state, the action/state depicted in this new pair is instead predicated of an angel. Given the angel’s status as God’s proxy, however, this conceptual extension may well be regarded as a natural analogy. With respect to the action/state of the angel represented, the two variant pictures are complementary in content. In P9, the angel at the gate is beholding the appearance of Eve in paradise by God’s creation (state). At the other

132 As will be shown in the following section, a more convincing explanation, which reinforces the synchronically based account, is forthcoming from a diachronic–comparative perspective.

240 

 11 The Expulsion: P45/P46

extremity, the angel at the gate is guarding it after the couple has disappeared from paradise by God’s expulsion (action). The right half of P46 (P46.2) associates with P13 through parallels and contrasts. In both, Adam and Eve are left on their own after having been addressed by God (in P45.1 and P10, respectively).133 But while they are protected in paradise in P13, as symbolized by the architectural frame surrounding them, they are situated in the outer world, an unframed space with no arched structure but the firmament above, as described in Genesis A 955a–957b: “ac hē him tō frōfre / lēt hwæðere forð wesan // hyrstedne hrōf / hālgum tunglum // and him grundwelan / ginne sealde” (but for their comfort he nevertheless let the sky continue, decorated with the holy stars, and gave them the broad bountiful land; Anlezark’s [2011: 71] translation). A further distinguishing feature concerns the arrangement of Adam and Eve. They are placed back to back in P46, whereas they face each other in P13. This reversal in configuration may symbolize the original perfect harmony between the couple—as suggested in conjunction with their identical pose—and the subsequent lack thereof. This network of association with P13 may offer a supplementary account of the change in relative position from P45.1/P45.2 (action) to P46.2 (state), observed above. P13 (state), too, implements a comparable reversal in positioning of the couple from P10 (action), whereby Adam moves from right to left relative to Eve, exactly the same alteration executed in P46.2. Since the distance to God is of no relevance in determining relative arrangement in the case of P13, and P13 arose well before P46, it may seem legitimate to assume that the repositioning of the couple in P46.2 would have been conditioned by the corresponding configuration in P13 because of their close association. It may be worth pointing out at this juncture that the concatenation of Angel– Adam–Eve (left to right) in P46 constitutes a reversal of the order of Eve–Adam–God (left to right) in P45.1. Although such a purely formal operation may derive the arrangement in P46, it remains unexplained why this change was implemented at all. No functionally motivated account seems forthcoming, in stark contrast with P31, for which the leftright reversal in P31.2 may adequately be characterized as an iconic representation of the cataclysm in human history (section 8.1). The allegedly corresponding reversal in P46, however, cannot be identified as analogous to that in P31, since the change in linear order does not constitute a prime compositional difference between P45.1 and P46: the two scenes do not feature the same characters with the angel replacing God, nor do the respective events take place in the same location. More specifically, inasmuch as this alleged conversion maps one of the two configurations involved into the other, it entails treating them as equivalent in cognitive terms as well. The assumed overall equivalence seems questionable, however: While there is no spatial break between God and Adam/ Eve in P45.1 because they are all located in paradise in direct confrontation with each other, the angel and the couple in P46 are situated in divergent spheres as symbolically

133 It may be recalled here that P45.1 is comparable to P10, as observed above.

11.2 A diachronic–comparative perspective 

 241

distinguished by the presence and absence of framing, as noted above. Moreover, while Adam looks God in the eye in P45.1, he does not seem to be gazing specifically at the angel. Rather, he is glancing back toward the gate of paradise as a whole. In this light, a simple mechanical reversal in defiance of the intervening segmentation in force and differing lines of sight would have been unlikely as a conceptually viable operation. The arrangement of Adam at left and Eve at right in P46.2 as determined above would have in turn singled out Adam—to the exclusion of Eve—to glance backwards to the angel because of the conceptual opposition to P13 noted in the preceding paragraph. Both Adam and Eve turning back alternatively would have detracted from the maximal contrast with the couple’s configuration in P13. Moreover, only Eve looking back, as realized in several cognate representations (see the following section) would have resulted in a face-to-face composition in P46.2 that would be indistinguishable from P13. All things considered, then, Adam exclusively came to be depicted turning back, as is actually the case in P46.2, despite the rarity of this option adopted elsewhere (for details, see the following section).

11.2 A diachronic–comparative perspective As comparison with other cognate works makes evident, the Expulsion scene in Junius (P45/P46) exhibits several rare or unique features. In order to substantiate the claim, one should begin with an examination of related pictures in the Cotton Genesis family and/or the Touronian Bibles. As practiced throughout this book, a structural analysis predicated on a selected set of distinctive features will offer a revealing overview. Drawing on previous studies, particularly Kessler (1977: 20–22) and Klein (1984: 95–98), one may devise the following set of distinctive features for a formal analysis of the wide variety of representations of the Expulsion manifested in the genealogical groups under consideration. 1. Clothed or naked: Adam and Eve are clothed or naked in the Expulsion. The clothed couple appears in Junius, Grandval (Pl. 24), Bamberg (Pl. 23), and San Marco (Pl. 51). Salerno (Pl. 57) is ambivalent, as the pair is half-clothed (only the lower body). Moreover, Vivian (Pl. 27) is less than clear: according to Kessler (1977: 20n28), the clothes are an addition. The remaining works—San Paolo (Pl. 34), Hildesheim (Kessler 1977: fig. 5), Millstatt (Pl. 66), the San Paolo frescoes (fol. 29r; https://digi.vatlib.it/view/ MSS_Barb.lat.4406/0030; Kessler 1977: fig. 31), and the Hortus Deliciarum (Pl. 69)— represent the couple naked in their departure from paradise. 2. Adam/Eve hold work tools when departing from the gate. Only in Junius and San Marco are instruments for farming (Adam) or weaving (Eve) carried by the couple. 3. Adam and/or Eve look back to the gate when leaving paradise. This parameter varies widely and without marked preference patterns. In Junius and Salerno, only Adam

242 

4.

5.

6.

7.

 11 The Expulsion: P45/P46

glances back. In contrast, only Eve turns back in Hildesheim, San Marco, Millstatt, and the San Paolo frescoes. Both Adam and Eve look back in Grandval, Bamberg (according to Kessler [1977: 20]), and the Hortus Deliciarum, whereas neither do so in Bamberg (according to Klein [1984: 95], which seems more plausible than Kessler’s reading given immediately above), Vivian, and San Paolo. At the gate, God or the angel drives the couple out. The vast majority take either of the options. God is shown pushing the couple in San Marco, Millstatt, and the Hortus Deliciarum, whereas an angel expels them in Grandval, Bamberg, Vivian, San Paolo, Hildesheim, the San Paolo frescoes, and Salerno.134 The sole exception is Junius, in which neither God nor an angel figures as expeller. This is a subcategory of 4 given above. Where an angel drives the couple out, it may hold a sword (Bamberg, San Paolo, Hildesheim, and the San Paolo frescoes) or not (Grandval, Vivian, and Salerno). In the absence of an angel as expeller, this feature is immaterial in Junius (where no expeller appears), San Marco, Millstatt, and the Hortus Deliciarum (in the latter three God serves as expeller). An angel guards the gate on its own, whirling a sword but not directly driving out the couple. By definition, where an angel expels the couple, this feature does not apply. Only three works realize this marked feature: Junius, Millstatt, and the Hortus Deliciarum. This feature is extraneous to the Expulsion scene per se. Rather, it concerns an autonomous scene of At Labor, which immediately follows the Expulsion more often than not. At issue is Eve’s specific work represented—nursing a child (Grandval, Bamberg, Vivian, San Paolo, Hildesheim, and the San Paolo frescoes), weaving (San Marco and the Hortus Deliciarum, Pl. 70), or tilling like Adam (Salerno, Pl. 57 right). Outstanding exceptions are Junius and Millstatt, which fail to represent a scene of the Labor itself. In addition, feature 2 (the couple holding something in the Expulsion) and feature 7 (Eve nursing or weaving) are mutually dependent in conceptual terms. Specifically, “feature 2 = Eve holding a tool for weaving” and “feature 7 = nursing” are at odds with each other. Therefore, at San Marco, where Adam and Eve are shown equipped with farming and weaving instruments, respectively, Eve is not nursing a child but weaving at the Labor scene.

The seven distinctive features that have been identified above as constituents of the Expulsion scene are distributed among the members of the Cotton Genesis family and the Touronian Bibles, as summarized in the following table.

134 In San Paolo, a wingless being with a nimbus swirls a sword. The yellow toga he wears as well as the sword he raises identifies him as an angel rather than God, who is consistently shown wearing a red toga.

11.2 A diachronic–comparative perspective 

 243

Table 13. Feature specifications of the Expulsion scene in the Cotton Genesis family members Feature Junius

G 

B 

V 

SP

H 

SM

M 

SPF

HD

Salerno

1  2  3  4  5  6  7 

c  no AE A  no —  n 

c  no no A  yes —  n 

c  no no A  no —  n 

n  no no A  yes —  n 

n  no E  A  yes —  n 

c  yes E  G  —  —  w 

n  no E  G  —  yes — 

n  no E  A  yes —  n 

n  no AE G  —  yes w 

(c) no A  A  no —  f 

c  yes A  —  —  yes — 

Key (to works; for details, see section 1.2): G = Grandval; B = Bamberg; V = Vivian; SP = San Paolo; H = Hildesheim; SM = San Marco; M = Millstatt; SPF = San Paolo fuori le Mura frescoes; HD = Hortus Deciliarum. Key (to features): feature 1, c = clothed, n = naked; feature 3, A =Adam, AE = Adam and Eve, E = Eve; feature 4, A = angel, G = God; feature 7, n = nursing a child, w = weaving, f = farming.

The comprehensive overview of the variation in the Expulsion provided above may put Junius in perspective among the cognate set and bring to light a body of unique or rare properties distinctively characteristic of P45/P46. First, Adam and Eve in the Expulsion hold certain tools for their labor (feature 2). In this respect, San Marco is the sole comparable example (Broderick 1978: 235–236). Unlike San Marco, however, Eve holds a whorl or disk—something not very conspicuous—in her hand rather than a spindle or distaff with greater visibility. Second, Adam, to the exclusion of Eve, looks back toward the gate of paradise (feature 3). Salerno offers the only instance apparently similar to it. Strictly speaking however, at Salerno Adam glances back to the angel who pushes him out (see the following point). Third, neither God nor an angel drives the couple out (feature 4). This is a unique property unknown elsewhere. Fourth, an angel is depicted guarding the gate as opposed to expelling the couple (feature 6). This feature is isolated, only common to Millstatt and the Hortus Deliciarum. Fifth, a Labor scene is absent, as only in Millstatt (feature 7). The value assignment to each of the above features is inherently biased and depends on textual specification in such a way that there is usually an unmarked value most likely to obtain, insofar as no overriding factors are at work that need to be specified. According to the biblical text, “feature 1 (clothed or naked) = clothed” (Gen. 3:21) and “feature 4 (God or angel driving out the couple) = God” (Gen. 3:23–24) are expected to be chosen. Accordingly, these default values may require no explanation. “Feature 2 (tools or none) = none” and “feature 7 (Eve nursing a child or weaving) = nursing” seem also unmarked, in the absence of textual determination for feature 2 (in contrast to clothing; see feature 1 above) and in light of Genesis 3:16 for nursing, respectively. By contrast, “feature 4 (God or angel driving out the couple) = angel” would have been a secondary development through merge with feature 6. That is, the occurrence of an angel as specified in “feature 6 (angel as guard) = yes” in accordance with the Book of Genesis (Gen. 3:24) would have been generalized to replace the original specification of God for

244 

 11 The Expulsion: P45/P46

feature 4 (driving out the couple; Broderick 1978: 241), and upon conflation with feature 4, feature 6 itself would have lost its original autonomy as an independent feature (indicated accordingly as ‘—’ in the above table). Thus, the angel would originally have been conceptualized as guard of the gate in line with the Book of Genesis, and a corresponding picture would have been prepared as a separate image. In this sense, the retention of “feature 6 (angel as guard) = yes” must be faithful to the original state of affairs. Finally, the scene of the Labor is most likely to follow the Expulsion, as is actually the case with the vast majority, given Genesis 3:23. Accordingly, the absence of a corresponding scene in Junius and Millstatt would be in need of explanation. From the above considerations, it may follow that Adam/Eve holding something in the Expulsion (feature 2), Adam/Eve leaving on their own without being expelled by God or an angel (feature 4), and the absence of a Labor scene (feature 7) must be characterized as innovations in Junius. The disuse of the Labor as an autonomous image (feature 7) would presumably have induced feature 2 by conflation, whereby the core information contained in the Labor scene—Adam tilling the ground (see the following paragraph)—was transferred to the Expulsion to be encoded in another form, namely, Adam holding farming implements. Millstatt, however, shows that the transference of information would hardly have been a necessary and mechanical consequence of dismissal of the Labor as an independent scene. The invariant presence of Adam tilling the ground in all Labor pictures, in contrast to the variation on Eve’s labor—nursing versus weaving—may testify to the primary importance of Adam’s activity. Also of interest in this regard is the inconsistency with which the spindle whorl is depicted in P45.2 (present) and P46.2 (absent): this contrast between consistency and lack thereof in representation may be ascribed to the difference in relative importance attached to Adam’s and Eve’s labor. This diminished visibility of Eve’s implement in Junius, compared with Adam’s counterpart in the same work and Eve’s possessions at San Marco and the Hortus Deliciarum, may be construed as the artist’s lack of confidence in its visualization, which in turn may be attributed to its less than firm and stable status as Eve’s symbol at the time of picturing this item. In view of the predominance of nursing a child as Eve’s labor (Table 13), and adopting a diachronic perspective, one may conjecture that the depiction of a whorl, a humble item of possession as it is, would have been a later (and perhaps still hesitant) addition when a secondary association between Eve and act of spinning became relatively known (further on this issue, see section 12.7.3 below). As argued above, the absence of the Labor scene in Junius would have served as a functional motivation for encoding labor equipment on the departing couple as a measure of compensation for keeping information that otherwise would have been lost upon erasure of the Labor scene. As a consequence, Junius is doubly removed from the Touronian Bibles with respect to features 2 (Adam and Eve holding work tools in their departure from paradise) and 7 (absence of the Labor scene). In contrast, San Marco (feature 2) and Millstatt (feature 7) are deviant only in one respect. As regards the treatment of features 2 and 7 in terms of information recovery, however, the maximally

11.2 A diachronic–comparative perspective 

 245

deviant Junius is the most efficient because it suffered no substantive information loss, in comparison with the vast majority cases with unmarked values for these features— the departing couple holding nothing and Eve nursing a child. San Marco is redundant, since the same information is replicated in two scenes, in the Expulsion and the Labor. And finally, Millstatt is the least efficient due to loss of information, because the original association of Adam with farming equipment has been lost irrecoverably without compensation implemented. With the effect of the erasure of the Labor scene in Junius having been understood, one is now confronted with a question in another direction: What was responsible for the absence of the Labor scene in Junius (and Millstatt, for that matter)? Were these two phenomena in the two manuscripts connected, or were they separate and only contingently similar? It might be conceivable to attribute the absence of a Labor scene in Junius to the double depiction of the Expulsion in P45/P46. According to Henderson (1975: 126) and Broderick (1978: 236), the doubling of the Expulsion scene led to a considerable lagging of picture behind text. A further depiction of the Labor would have exceeded a level of tolerance—whatever it may be in precise terms—for misalignment between text and image. Rather than by appealing to such a vague extrapictorial notion of verbal–visual discrepancy, an immanent interpictorial account would seem worth exploring, under the assumption that the disuse of the Labor as an autonomous image in Junius was a structurally induced innovation. Specifically, it would have been ultimately triggered by the creation of the couple’s departure scene as a direct sequel to God’s final pronouncement (P45.1), that is, P45.2, which was internally inspired, directly by P45.1 and generally by the core network of interpictorial relations in which the novel representation (P45.2) was firmly placed as a fully integrated image, as detailed below. This innovation realized the couple’s departure at the gate prior to and independent of the angel previously depicted as their expeller. This reordering of the two scenes turned out to be in full conformity with Genesis A 945a–947b: “him on laste belēac / līðsa and wynna // hihtfulne hām  / hālig engel  // be frēan hǣse  / fȳrene sweorde” (Behind them, at the Lord’s behest, a holy angel with a fiery sword closed the joy-filled home of leisure and pleasure; Anlezark’s [2011: 71] translation). The compositional dissociation of the angel from Adam and Eve was conceptually equivalent to the abandonment of the motif of the angel expels the couple from paradise. Accordingly, the figure of the angel standing alone newly obtained reconstituted the scene of the angel guarding the gate (see above). That is, the angel as expeller, now detached from the expelled couple conceptually and compositionally, was reconceptualized (or reaffirmed, from a diachronic perspective; see below) as guardian of the gate standing on its own. This in turn would have securely repositioned the couple—depicted being driven out by the angel in the Touronian Bibles—in the field away from the gate, thereby making the resultant scene approximate to the Labor scene immediately following the Expulsion in the Touronian Bibles. That is, the emergent right side of P46 (P46.2), representing the open field as background rather than the gate of paradise as in some other cognate works (Hildesheim, San Marco,

246 

 11 The Expulsion: P45/P46

and Millstatt), would have been found compatible with interpreting it as an implicit depiction of the Labor scene. Given the exemplary status of the Touronian Bibles for the Junius artist, this approximation to the model image of the Labor would have been appreciated as capable of serving as a substitute for an explicit representation of the scene in question. Thus, the Labor scene failed to be reproduced as an independent picture in Junius; instead, it was absorbed into the Expulsion series as its final phase through a merge with the earlier departure scene as reflected in the Touronian Bibles, itself now radically recomposed. The foregoing formal–structural account needs substantiation for its empirical credibility. In executing reconfiguration and reconceptualization in Junius, the artist seems to have built on Bamberg 7v3 (right), in light of the virtual pictorial identity involved between P46 and Bamberg that Raw (1953: 58–59; 1976: 142) draws attention to on substantive grounds: “The Junius drawing is particularly close to the representation in the Bamberg Bible, as regards the stance of the angel, the position of the sword and even the floating panel of drapery over the angel’s right arm” (Raw 1976: 142).135 In reworking the Bamberg model, the Junius artist literally split the configuration of the angel and the couple into two autonomous subcompositions, the angel standing on its own as guard on the left side (P46.1), and the couple departed from the gate to the open field on the right side (P46.2). Upon executing such a bisection of the unitary configuration in the model, the artist introduced further distinctions in the surroundings of the resultant two configurations in accordance with the new motif reconstituted. Among the many modifications implemented are the form and decoration of the gate on the left side in similitude to the comparable entity in P9, and the depiction of the sky and the ground with plants on the right side in conformity with the passage, Genesis A 955a–961b: ac hē him tō frōfre  / lēt hwæðere forð wesan  // hyrstedne hrōf  / hālgum tunglum  // and him grundwelan / ginne sealde. // hēt þām sinhīwum / sǣs and eorðan // tūddortēondra / teohha ge­­ hwilcre  // tō woruldnȳtte  / wæstmas fēdan.  // Gesǣton þā æfter synne  / sorgfulre land (but for their comfort he nevertheless let the sky continue, decorated with the holy stars, and gave them the broad bountiful land; he commanded each fecund species of sea and earth to offer fruits to the married pair for their worldly need. Then after the sin they settled the sorrowful land,  …) (Anlezark’s [2001: 71, 73] translation)

Thus, in reconstituting the Expulsion scene, the Junius artist used the Touronian Bible— as direct material to improve on for a radical recomposition—as well as the Cotton Genesis model—as an overall framework to draw on, especially with respect to the restoration of the scene of an angel guarding the gate of paradise. The net result of the artist’s reworking of features 4 and 6 is that Junius reverted close to the Cotton Genesis 135 Broderick’s (1978: 241) criticism of Raw’s assessment is insubstantial. He simply calls into question Raw’s privileging of the Bamberg Bible among the cognate works that are all subject to conflations in their own ways. Raw’s point is that, despite superficial alterations, the Bamberg and Junius figures are highly analogous, a contention that bears empirical scrutiny.

11.2 A diachronic–comparative perspective 

 247

original by undoing the conflated representation of the angel driving out the couple— the practice well established in the Touronian Bibles—and thus recovering the image of an angel guarding the gate of paradise. As reasoned above, in contrast to the innovative features 2, 4, and 7, feature 6 (angel as guard) would have been a reflex of the original state of affairs in the Cotton Genesis tradition. Yet the attestation of this property in Junius must be understood as a pictorial innovation, activated, as expounded above, by the artist’s rejection of the otherwise prevalent practice in the Touronian Bibles of merging features 4 and 6 (whereby feature 4 = angel, not God as in the original; cf. Broderick 1978: 234–235, 241). The convention at issue was in force largely due to the space limitation imposed by the register-based single-page format for the Touronian Bibles. This imaginative treatment of feature 6 is a rare outstanding example of Junius not conforming to the Touronian Bibles. One may naturally wonder why this Touronian practice was abandoned, while many other examples and conventions were adopted faithfully. A partial answer to this question has already been alluded to above. Not intended to be a part of a one-volume Bible (pandect) but a Genesis book standing on its own or more probably a selection of episodes from the Book of Genesis, the limitation of a single-page miniature no longer applied to the Old Saxon Genesis preceding the Old English/Old Saxon Genesis. Thus, the Junius artist would have been freed from the original dictation for space saving that the illuminators of the Touronian Bibles had been subject to. A deeper-rooted reason, however, would have been the internal logic of the pictorial organization of Junius in parallel to P3.2b/ P3.3 and P16.1/P16.2, as argued above and further substantiated below. By way of recapitulation, elaboration, and integration, the chain reactions eventually leading to the restoration of the angel guarding the gate of paradise and the abandonment of the Labor scene started with the Judgment scene, P44. As a realization of symmetrical representation, albeit nonprototypical in manifestation (section 10.1), P44 would have inspired composition of its asymmetrical counterpart—God’s final pronouncement for expulsion—in line with the recurrent pattern that permeates the system of the pictures provided so far in Junius (section 10.1). As a consequence, P45.1 came into being. As a reminder, this picture, unique to Junius, has no correspondents in cognate works. As is often the case with the asymmetrical/profile variants (e.g., P2), the lower half is left blank, which may be filled by another scene if necessary. By virtue of its inherent dynamicity, the action visualized in the asymmetrical variant in general usually gives rise to an event as its consequence, as with P3.2b resulting in P3.3. In parallel fashion, a prime result consequent on P45.1 would be most likely to be accommodated in the open space of page 45. Accordingly, a scene of Adam and Eve’s departure from the gate materialized as P45.2, with a concomitant reinterpretation that God expelled the couple from paradise, as inferred from P45.1. Once actually put to use as the Expulsion (P45.2), this structurally motivated space would in turn have induced a splitting of the conflated representation of the angel driving out Adam and Eve, the motif that had been firmly established in the Touronian Bibles. A simple replication of the Touronian scene—the angel pushing the couple out

248 

 11 The Expulsion: P45/P46

the gate—would have rendered itself redundant at best and contradictory at worst. The rejection of the Touronian convention that the artist—in all likelihood, a member of the Tours School; section 12.8—was induced to implement on internal grounds would have been such a radical disruption of the pictorial tradition that it would have necessitated overriding authentication for execution. The needed sanction would have been provided by the Cotton Genesis model that was concurrently available for consultation, in conjunction with the textual authority. At this point, the angel as keeper of the gate of paradise assumes paramount significance. In accordance with Genesis 3:24, the Cotton tradition accords an independent status to a corresponding image, as reflected in Millstatt and the Hortus Deliciarum, although in the Touronian branch the scene has long been deprived of its pictorial autonomy through conflation. Thus, endorsed anew by the Cotton Genesis exemplar with the biblical authenticity, as well as by the corresponding Old Saxon antecedent reflected in the Old English/Old Saxon Genesis passage (945a– 947b), the Junius artist would have confidently restored the scene of the angel guarding the gate and reinterpreted the couple being driven out from the gate as standing far removed from it in the outer world. Having thus dissociated the couple from the expelling angel conceptually and distanced them from the gate physically, the artist would have perceived the resultant image in the open field as a significant approximation to the Labor scene. Because of the appreciable and desirable conceptual and compositional overlap that the artist became convinced of recognizing between the two scenes, he would have decided on merging the two into a single picture, P46.2, and done away with the Labor scene proper. Thus, the Labor scene ended up being absorbed into P46.2 to lose its independence as visualization. A complete suppression of the Labor scene, however, was avoided: part of the information contained there was retained and transferred to P45.2 and P46.2 as farming tools Adam holds.136 Also noteworthy in this connection is the hilly ground in P46.2, which is reminiscent of the Labor scene in the Touronian Bibles in general and the Bamberg Bible in particular. This correspondence can be regarded as a further vestige of the underlying Labor scene conflated with the departing couple in Junius. On a final note, the absence of the Labor scene in Junius and Millstatt must be regarded as independent of each other. Apparently the same phenomena diverge both in their cause and effect. In contrast to Junius, the loss of the scene (feature 7) did not incur any compensation measures in Millstatt, especially with respect to the couple’s possessions in the Expulsion (feature 2). Nor can the common retention of the angel as guard of the gate (feature 6) be held equally responsible for the disuse of the Labor as an autonomous scene in light of the Hortus Deliciarum. Most significantly, the fundamen136 While Henderson (1975: 127; see also Broderick 1978: 234–235) refers to an emotional aspect of the avoidance of depicting God as expeller (“It may be that the illustrator was uneasy about the comparatively rare expulsion-by-God imagery, . . .”), I prefer to emphasize the cognitive and rational basis of the artist’s decision on this matter.

11.2 A diachronic–comparative perspective 

 249

tal cause for the loss of the Labor scene in Junius—the creation of P45—did not obtain in Millstatt, as the system of pictorial organization in Junius was evidently foreign to Millstatt. All things considered, then, the ultimate cause for the defining features 2, 4, 6, and 7 of Junius would have been one and the same, the internally inspired composition of P45 as paired with P44, in line with P3.2b/P3.3, P6/P7 (see section 12.3), P10/11/13, and P16.1/P16.2.

12 The Touronian Foundations of the Old English Genesis: A Synthesis 12.1 A classification of primary models: the four major sources A primary model (or models, as a complex of models is often involved) of a given picture may be defined as material that served as a major source of inspiration and influence for its composition in terms of the overall pictorial organization and specifics of individual figures and their configurations in the entire space of visualization, rather than restricted to borrowing of isolated entities and their arrangement within a limited space. Since more than one source of inspiration may be involved in competition and/or complementation, it is often feasible to address multiple primary models. To distinguish between primary models in multiples and a multiplicity of ranked models is obviously an empirical issue that must be figured out specifically on an individual basis. The twenty-two Junius pictures under investigation, and also their smaller constituent pictures in some cases, are divided into four classes, on the basis of the identity of their primary sources as follows: (i) internal sources of inspiration, that is, other Junius pictures or the system of pictorial organization underlying them; (ii) the part of the Touronian Bibles not derived from the Cotton Genesis; (iii) the part of the Touronian Bibles that has its origin in the Cotton Genesis; and (iv) Cotton Genesis sources that were not incorporated into the Touronian Bibles. Under class 1, more often than not it is an abstract structural relation, rather than a concrete pictorial manifestation, that provides a prime internal basis for construction of another picture. Given that the underlying system and its embodiments through concrete pictures are mutually dependent and thus inseparable in terms of their actualization in the empirical world, it is often difficult to make a categorical distinction between the two notions. These four classes of primary models may be formalized most economically and revealingly in terms of the pairing of two binary parameters, Touronian/Non-Touronian (or ± Touronian) on the one hand and Cottonian/Non-Cottonian (or ± Cottonian) on the other, as presented in the following: Table 14. Classification of primary sources for Junius pictures Feature

Cottonian

Non-Cottonian

Touronian Non-Touronian

Class 3 Class 4

Class 2 Class 1

https://doi.org/10.1515/9783110788068-012

252 

 12 The Touronian Foundations of the Old English Genesis: A Synthesis

By way of summary and elaboration, each category is fully exemplified in the following. In the many cases where multiple sources must be postulated as primary models—competing or complementary—they are all enumerated without determining their relative importance among them. Similarly, where multiple ranked models are specifically at issue, they are noted as such. Table 15. The major sources of the first twenty-two pictures (including their constituent pictures) in Junius 11 Class 1. Non-Touronian/Non-Cottonian (internal) sources Picture

Episode

Model

P11

God beholds Adam and Eve in paradise

P13 P16.1 P16.2 P17.1 P17.2 P20.2 P34.1 P36.1 P36.2 P39.1 P39.2 P41.2

Adam and Eve in paradise God rebukes the rebel angels Hell God beholds the fallen angels in hell Hell Hell Covering Covering Hell Covering Covering God addresses Adam and Eve, who hide from him God pronounces expulsion Expulsion

P10, by analogy to Pii/P2; secondary models: for animals, Bamberg 7v1 and Millstatt 9r; for God framed, Vivian 329v, 423r, and San Paolo 259v P11; secondary model: Millstatt 8r P2, P3.2b P3.3 Pii P3.3, P16.2; by analogy to Pii/P2 P3.3, P16.2, P17.2 P34.2 P34.2/P34.1 P3.3, P16.2, P17.2, P20.2 P34.2/P34.1/P36.1 P34.2/P34.1/P36.1/P39.1 P31.2, by analogy to P10/P13

P45.1 P45.2

P10, P44 P45.1, by analogy to P2/P3, P3.2b/P3.3, P10/P11/ P13/P20.1, P16.1/P16.2

Class 2. Touronian/Non-Cottonian sources Picture

Episode

Model

Pii

God enthroned

P2

God enthroned

P3

Fall of the rebel angels

Vivian 423r; secondary model: Touronian image of Christ in majesty Vivian 423r; secondary model: Touronian image of Christ in majesty Vivian 3v, 386v, San Paolo 3v, 310v

12.2 The internal models (class 1) and the system of pictorial organization 

 253

Table 15. (continued) Class 3. Touronian/Cottonian sources Picture

Episode

Model

P9

Creation of Eve

P10

God welcomes Adam and Eve in paradise Temptation of Adam

Vivian 10v1, Bamberg 7v2 (mirror image); ­Creation of Adam + (Observing Angel) + Drawing a Rib; by opposition to P20 Grandval 5v2

P20.1 P24 P28 P31.1 P31.2 (P41.2) P34.2

P41.1/P44 P46

Grandval 5v2 (right), 5v3 (far left); Admonition + Temptation of Eve Temptation of Adam San Paolo 8v1; San Marco; Animation of Adam Temptation of Adam and Eve Vivian 10v2, Bamberg 7v2 (mid); Temptation of Adam and Eve (snake + Eve + Adam) Fall of man Grandval/Vivian/Bamberg/San Paolo/San Marco; Introduction of Eve to Adam Adam and Eve pray and lament Bamberg 7v2 (right), mirror image; Covering/ Hiding + Eve spreading her arms Adam and Eve cover themselves Symmetrical variant of Bamberg 7v2 (right), such as HD and SP; Covering/Hiding, often conflated with other scenes Judgment Bamberg 7v3 (left; partial mirror), Grandval 5v3 (right; partial mirror); Reproval/Denial Angel guards the gate and Adam Bamberg 7v3 (right) and Eve expelled Class 4. Non-Touronian/Cottonian sources

Picture

Episode

Model

P6

Creation

P7

Creation

Salerno; partly King on the throne, the Evangelist at work, canon tables such as Vivian 326r Salerno; partly King on the throne, the Evangelist at work, canon tables such as Vivian 326r

12.2 The internal models (class 1) and the system of pictorial organization The first class—internal sources—mainly concerns composition of novel scenes that seem to have no correspondences in the Cotton Genesis, as reconstructed in Weitzmann/ Kessler (1986), or in the Touronian Bibles. These are innovative scenes that are unknown in the biblical tradition, as epitomized in the Cotton Genesis and its family. It is not necessarily true, however, that the novel visualizations would have been necessitated by the distinct stories characteristic of the Old English/Old Saxon Genesis. Rather than required by narrative contents, many of these new images were motivated, structurally

254 

 12 The Touronian Foundations of the Old English Genesis: A Synthesis

and autonomously, by the underlying system of pictorial organization in Junius: they would have been created primarily out of intrinsic motivations of visualization through opposition to, contrast with, and variation of extant pictures according to the set of rules and principles that the Junius artist followed with fair consistency and strictness. While of paramount significance in exploring the cognitive bases of image creation, this class does not figure centrally in addressing genealogical issues. Therefore, in reconstructing genealogical relations of Junius, one must exclude from consideration the pictures that would have been motivated primarily on internal grounds. The internal derivation, however, presents its own problematics to be treated squarely below. The construction of new pictures based on internal models may be divided into two groups, depending on whether an emergent image constitutes a novel scene or simply a variant of an existing representation without conveying new information in terms of narrative contents. Under the first group fall P11, P13, P16.1, P17.1, P17.2, P20.2, P36.2, P41.2, P45.1, and P45.2. The second group is constituted by all variants of P34.2 (Covering), namely, P34.1, P36.1, P39.1, and P39.2. Since the second group is one of narrative redundancy, the model may be readily specified, once a redundancy-based variation has been executed. By contrast, given that the first group is characterized by open-ended creativity, the range of models may be regarded as correspondingly open. One might be tempted to expect that any earlier picture may be chosen as a model for a subsequent image insofar as favorable conditions obtain. Nothing could be further from the truth than this prediction. As it turns out, it is far from true that any earlier picture in Junius may equally inspire a subsequent one. On the contrary, only a few privileged ones are qualified for such internal stimulation. Two of the three picture cycles represented in the exclusive set of the twenty-two pictures under investigation (section 1.1), namely, the Fall of the Rebel Angels cycle and the Adam and Eve cycle, exemplify internally inspired creation. The Fall cycle provide two such internally motivated images of God (P16.1 and P17.1, derived from P2 and Pii, respectively) and two further of hell (P16.2 and P17.2, both derived from P3.3). Since the major occupant of hell is no other than Satan, the two pictures of hell may be identified as representations of Satan (see below for further consideration). As for the Adam/Eve cycle excluding, for the reason adduced above, the variation on the Covering (P34.1, P36.1, P39.1, and P39.2), one may recognize four God images (P11, P13, P41.2, and P45.1) internally derived and two representations of hell (or Satan) comparably composed (P20.2 and P36.2). A further instantiation is an internally motivated image of Adam and Eve (P45.2), which is itself derived from an internal derivative featuring God, that is, P45.1. As listed in Table 15 above, while the pictures of hell/Satan are based on P3.3 (and all succeeding varieties), all the remainder, in the final analysis, draw on Pii/P2, directly or through the intermediary of P10 and/or P11, a derivative pair of Pii/P2. Given that Satan, God’s antagonist par excellence, is defined in opposition to God (see below), one can readily appreciate that the figuration of God at large has an overwhelming privilege of serving as an internal source for image proliferation. It is not immediately evident, however, how the particular pair, Pii/P2, among the wealth of God

12.2 The internal models (class 1) and the system of pictorial organization 

 255

images, assumes a monopoly over image production. Thus, one encounters a vital question at this point: Why does the pair of Pii/P2 monopolize such a generative capacity for image proliferation? Undoubtedly, appearing earlier than any other picture in the manuscript, these two pictures have the advantage of setting the stage, but their precedence in pictorial narration cannot fully substitute for a principled explanation. In order to offer an explanatory account for the question, one needs to determine how the pictures of God and related beings including Satan are internally organized in Junius. Only with reference to the whole perspective of organization can one afford to appreciate the privileged status of Pii/P2. As substantiated throughout the preceding chapters, the Junius pictures are implicated in the dense structural network of parallelism, contrast, and variation. Specifically, it should be clear by now that the representation of God, his associates, and his enemies is far from a contingent inventory of idiosyncratic images constructed on an ad hoc basis. Drawing on the individual examinations conducted so far, it may be appropriate here to provide a comprehensive overview of the network of pictures by way of recapitulation and elaboration. As mentioned from time to time in the preceding discussion, the whole constellation of God-related images actualized is formally organized on the basis of a few parameters, each of which is binary in value specification: (i) the types of figures (positive vs. negative; major vs. minor); (ii) the representation modes (detailed vs. reduced); and (iii) the types of scenes (state/event, denoted by a one-place predicate, vs. action, denoted by a two-place predicate).137 Maximally four distinct figures are involved in the system: God, Michael, Lucifer/ Satan, and Satan’s agent (unnamed). The pair of Lucifer/Satan are positional variants of the same identity, situated in heaven on the one hand and hell on the other. Unlike other extraterrestrial beings (angels and fiends), these four are identified by their unique roles in the narrative, and thus warrant individuated representation. However, they are far from equally subject to visualization. God receives most intensive and extensive representation in multiple modes and from varying perspectives. While depicted recurrently (and in fact, more frequently than God), Lucifer/Satan is represented from limited perspectives. Finally, the minor figures on both sides—Michael on the one hand and Satan’s messenger on the other—appear in only one of the three cycles under investigation, namely, the Adam and Eve cycle. Based on the differing extent of visualiza137 With particular reference to the changing representations of the episode of Moses raising his hands to seek victory at the battle with Amalek (Exodus 17:9–13) from late Antiquity to the early modern period, Schapiro (1973) originally brought to light the opposition between frontal and profile view, in conjunction with its corresponding conceptual contrast between state and action, explored their symbolism and further correlations in figuration and composition, and identified their underlying polar attitudes to and conceptualizations of life and art. While revealing the cross-cultural generality of the clustering of oppositions based on frontal/profile and state/action, Schapiro (1973: 42–49) correctly appreciated their varying manifestations and meanings in terms of culture and style specificity. In this light, the pictorial organization of the Old English/Old Saxon Genesis may be understood as a particular instantiation and systematization of the fundamental cognitive opposition.

256 

 12 The Touronian Foundations of the Old English Genesis: A Synthesis

tion conditioned by their varying participation in the narrative, as well as their mutual relationship, the four figures are bifurcated into two groups, positive (God’s side) and negative (God’s enemy), with a subdivision into two characters, major (God and Lucifer/ Satan) and minor (Michael and Satan’s agent), corresponding to the superordinate/subordinate dichotomy on each side. Two modes of representation may be contrasted, detailed and reduced. Primarily referring to the manner in which the four figures are depicted, this is largely a relative notion whose categorization is a matter of comparison. Where there are no definite indications in evidence for differentiation, one must acknowledge that the distinction between the two modes is neutralized in that particular context. Generally, the detailed mode is at work where a large space is allotted to visualization on its own. Accordingly, a full figure is depicted, with correspondingly careful attention also paid to the surroundings, such as architectural frames. The scenes depicted fall into two types, on the one hand, static (typically representing a state in which a given entity is found, that is, a theme is involved in a state/event, as denoted by an intransitive verb; see below), and on the other, dynamic (typically representing an action delivered by an agent on another entity, that is, an agent acts on a patient or theme, as denoted by a transitive verb). Viewed from the perspective of the figures, they are divided into two groups according to the thematic roles they play in the scene represented: the agent on the one hand and the theme (or patient) on the other. As denoted by a one-place predicate (e.g., be, fall), a theme is involved in a state or event without affecting another entity. By contrast, typically realized as grammatical subject of a two-place predicate (e.g., attack, deceive), an agent carries out an action on another being, usually treated as grammatical object. The static/intransitive scene is symmetrically constructed, while the dynamic/transitive one tends to be asymmetrically composed, as the agent and the recipient of an action—the theme or the patient—must be conceptualized as polar opposites—affecting versus being affected—and visualized with corresponding differentiation. With regard to the Fall of the Rebel Angels cycle, the following system of organization (Table 16) may be postulated underlying the wealth of visualizations materialized in the manuscript. Table 16. The pictorial organization of the Fall of the Rebel Angels cycle with respect to representations of God and related (extraterrestrial) figures Figure

Mode of representation

Positive

God

Negative

Lucifer/Satan

Detailed Reduced Reduced Detailed

Type of scene State/event (Pii) P17.1

Action

(P2) P3.2b, P16.1 P3.2a P3.3, P16.2, P17.2 P3.1

12.2 The internal models (class 1) and the system of pictorial organization 

 257

Given that the reduced mode is originally derived from the detailed one through abbreviation, and given that Lucifer/Satan is identified negatively in opposition to God, rather than vice versa, the foundational representations are embodied by the pair of Pii/P2, which are fully complementary in their manner and perspectives of visualization (section 2.1), a symmetrical representation as state (Pii) versus an asymmetrical representation as action (P2). Connected with such a foundational status of Pii and P2 formally determined in the above paragraph is their relative independence from the Fall cycle. Obviously designed as a frontispiece of the entire volume, Pii may be detached from a particular pictorial cycle. The treatment of P2 seems somewhat ambivalent. Insofar as a thematic continuity with P3 is recognized, it may be incorporated into the Fall cycle as one of its member pictures. As a depiction of the original and ideal state of affairs in which Lucifer is fully obedient to God and God favors him above all angels, P2 may be viewed as on a par with Pii, that is, dissociated from and prefacing the Fall cycle. One may be reminded in this regard of the inscription accompanying P2, “hǣlendes hēhseld” (the throne of the Savior; section 2.1). Giving full justice to this legend, one may be justified in assuming, as defended in section 2.1 above, that P2 serves as a complementary image of Pii. On this conceptual basis, these two initial pictures are treated as independent of and commonly underlying the three specific cycles to follow (the Fall of the Rebel Angels cycle, the Creation cycle, and the Adam and Eve cycle; section 1.1), as marked explicitly by means of parentheses in the above table (Table 16). It should be recognized that the opposition between positive (God) and negative (Lucifer/Satan) is uneven and asymmetrical, rather than even and complementary. As remarked above, the positive figure (God) is central and the negative one peripheral, as determined in conceptual terms, God versus God’s antagonist, rather than Satan versus Satan’s antagonist. This existential and epistemological asymmetry between positive (God) and negative (God’s enemy) is corroborated by the distinct bases of organization and their consequences for the differing practices of representation. First, while the positive domain is fully organized into four contrasting members by the maximal intersection of the two parameters of “state/event versus action” and “detailed versus reduced,” the negative domain is only partially structured, with neutralization in effect between state/event and action for the reduced mode. Due to this fully symmetrical organization of God images in contrast to those of Lucifer/Satan, the corresponding images are available without a gap between detailed and reduced, and between state/event and action. In the Lucifer/Satan domain, however, analogous perfect correspondences are not obtained because of its deficient organization. Second, while a full-page illustration may be provided for God (e.g., Pii), a half-page one is the maximum for Satan, even in a detailed version of its representation (P20.2, the Adam and Eve cycle). Third, and related to the second point, Satan is never represented in his own right as the sole figure in a picture, in contrast to God: Satan is always depicted as part of hell, inseparably in conjunction with his surroundings, and enclosed and packed together with a crowd of fiends. Finally, while God serves as an agent of action directed to Lucifer/Satan, the con-

258 

 12 The Touronian Foundations of the Old English Genesis: A Synthesis

trary never takes place. That is, wherever God and Lucifer/Satan are involved in action in a common scene, God invariably plays the role of an agent, whereas Lucifer/Satan is treated as a patient that is affected by God’s action. Against the background of the pictorial organization of the Fall of the Rebel Angels cycle (Table 16), the internal derivations identified earlier (class 1, Table 15) may be characterized as mappings from one given category to another. The base category and the derived category may differ (cross-categorial derivation) or remain constant (intra-categorial derivation). Depending on the status of categories involved, a given derivation may be characterized in specific terms, and can vary widely in its attending visualization effects, ranging from creation of a novel image hitherto unknown to a more or less redundant variation of an existing picture. In more specific analytical terms, the construction of P16.1 on the one hand and P17.1 on the other, based on P2 and Pii, respectively, may be commonly viewed as a reduction, since a member of the reduced category (P16.1 and P17.1) is created from one of the detailed one (P2 and Pii). The derivation thus proceeds from detailed to reduced in its path of image production. P16.1, however, is also modeled after P3.2b, a member of the same category in which P16.1 is situated. In this light, the derivation of P16.1 also involves an intra-categorial variation as its moment of creation. The three variants of Satan images, P3.3, P16.2, and P17.2, constitute a variation within the same category (negative, Lucifer/Satan, detailed, state/event). Inasmuch as no particular external model can be identified, the chain of derivation must be regarded as closed without further linkage to the outside. By contrast, the member of the minimally opposed category (negative, Lucifer/Satan, detailed, action), P3.1, is connected with an external model, Vivian 3v, 386v, and San Paolo 3v, 310v, as detailed in section 3.2 above. While manifested in larger numbers and attended by more participants than in the Fall of the Rebel Angels cycle, the representations of God, his associate, and his antagonists in the Adam and Eve cycle (Table 17) are executed with a diminished range of variation, whereby the contrast in the modes of representation (detailed vs. reduced) is totally neutralized, except for the core figures of God. Much the same applies to the opposition between state/event and action, which is retained only for the positive, major figure, God, and the negative, minor one, Satan’s messenger. With Michael and Satan’s emissary coming onto the stage, however, the contrast between major and minor figures is fully implemented in the entire system. Noteworthy is the reversal between major and minor when it comes to the contrast between state/event and action. While God (in the detailed mode of representation) is depicted either as a theme or an agent, it is not Satan, but Satan’s messenger, that exhibits a comparable opposition. This is subject to understanding, given that in the Adam and Eve cycle, Satan’s agent figures as a main actor in the narrative, as it significantly interacts with the couple on behalf of Satan, whereas Satan stays bound within the enclosure of hell. Pivotal status is accorded to the fully contrasted pair of P10/P11 in the Adam and Eve cycle, analogous to Pii/P2 in the Fall of the Rebel Angels counterpart, although the latter two make their presence felt as foundational images independent of specific cycles. The

12.2 The internal models (class 1) and the system of pictorial organization 

 259

other members of the detailed mode are not involved in a comparably systematized complementariy (Chapters 6, 10, and 11). According to the system of organization for the Adam and Eve cycle (Table 17), the internal derivation of P11 from P10 constitutes a cross-categorial operation, from action to state. A comparable cross-categorial analogy, but implemented in a different direction (from detailed to reduced), is realized by P13, which was inspired by its detailed counterparts, P10 (action) and P11 (state). The double variants of Satan, P20.2 and P36.2, may be simply characterized as an intra-categorial variation, both model and result belonging to the same category (negative, major). While varying within a single category as with the group of the Covering treated below, the variation between P20.2 and P36.2 concerns a category that is neutral to and hence unaffected by the dichotomy of detailed and reduced. Accordingly, the resulting variants, P20.2 and P36.2, do not suffer the same kind of conceptual redundancy as permeates the Covering group that is grounded on proliferation within a category defined as reduced.138 P34.1, P34.2, P36.1, P39.1, and P39.2 do not contain a figure of God or other extraterrestrial beings. Accordingly, one might be tempted to exclude them from the network of pictures in the Adam and Eve cycle as one addresses it from the perspective of the figuration of God and God’s antagonists. In the presence of P13, in which God is represented in an extremely tiny form, it seems logically consistent to generalize that these pictures have God represented in the absolute minimum, namely, in zero form. Somewhat analogous is P31.2, in which Adam and Eve pray to God in absentia in paradise. In the absence of God, however, Satan’s agent is depicted as a bystander on the scene.139 Similarly, the couple cover their nakedness in P34.1, P34.2, P36.1, P39.1, and P39.2, while God is away from paradise. In these instances, a demonic being does not make an appearance in lieu of God, however. Thus, the zero representation of God may be viewed as the most extreme form of reduction, even without a figural substitution. Analyzing P31.2 as a zero representation of God guarantees a formally rigorous characterization of it as a reduced variant of P41.2, the account that was intuitively followed in section 10.1 above. Correspondingly, the internally motivated derivation of P41.2 from P31.2 may be identified as a compositional elaboration predicated on the underlying pictorial organization in general and the opposition between detailed and reduced in particular. Inasmuch as a new picture is normally constructed on the basis of the founding member(s) represented in the detailed mode (as with the derivation of P13 from P11/P10), this case may be characterized as a reverse course of innovation, what linguists call back-formation, that is, the derivation of a basic (unmarked) form from 138 In addition to being reduced in the mode of representation, the set of the Covering pictures has reduction maximally implemented, so much so that the resultant zero figuration of God—see the following paragraph in the text—brings about bleaching of pictorial contents and militates against construction of new meanings. 139 Accordingly, P31.2 is included in two different places in Table 17, as it represents God in zero form but Satan’s agent in its explicit form.

260 

 12 The Touronian Foundations of the Old English Genesis: A Synthesis

a derived (marked) counterpart. In any case, regardless of directionality in analogical formation, this is another telling example of the generative power inherent in the organization system. Not simply a static description of the extant state of affairs, it should be conceptualized as a system equipped with creative dynamism. Analogously, the multiplication based on P34.2 may be regarded as a proliferation internal to a single category (positive, major, reduced), rather than crossing category boundaries, as with the derivation of P13 from P11 or P41.2 from P31.2 treated above. Inasmuch as the multiplication occurs on a reduced variant (the least specified category), the process—an intra-categorial variation—may be characterized as a variation at the most superficial level. By contrast, a cross-categorial derivation, the construction of a new image crossing category boundaries, is an operation on a deeper level that leads to a novel image of richer contents and greater potential in signification. One may well compare the variant group of P34.1, P34.2, P36.1, P39.1, and P39.2 with the contrastive pair of P10/P11 (action vs. state) or that of P31.2/P41.2 (reduced vs. detailed), to appreciate more profound effects exerted by cross-categorial derivation. In similar fashion, P45.2 may be identified as a reduced variant of P45.1: while God is depicted as an agent in P45.1, making a final pronouncement of expulsion to Adam and Eve, the figure of God is maximally reduced to zero in P45.2, so that the couple is shown departing paradise of their own accord. It may be of interest to compare P45.2 with San Marco (Pl. 51) and Millstatt 14v (Pl. 66), in which God drives the couple out of paradise. In P45.2, an explicit representation of God ends up being completely suppressed for the reason provided in section 11.1. Thus, P45.2 is derived through reduction (from detailed to reduced), as is P13 from P11, and contrary to the composition of P41.2 (detailed) based on P31.2 (reduced), a case of elaboration. The reduction, however, was executed maximally, resulting in a total erasure of the figure of God.140 As for P45.1, insofar as P10 serves as an internal model based on conceptual opposition, its emergence is owing to an intra-categorial variation. Since the category involved is the detailed mode and thus rich in potential for imaginative representation, the resultant image (P45.1) is capable of picturing a conceptually new scene far removed from that of the model (P10), in contrast to the intra-categorial variation of the Covering, which may give the impression of being more redundant than expressive of new meanings.

140 Another case of the reduction to zero is exemplified by P20.1 (right, section 7.2). While primarily based on the Admonition in Grandval 5v, one of the derivational processes responsible for it, namely, the removal of God from depiction, may be accounted for by identifying P10 as an auxiliary internal model for the zero representation of God. Correspondingly, P20.1 (right) may be characterized, synchronically in terms of the pictorial organization system, as an exaggerated variant of P13, distinguished as it is by the maximal mode of reduction.

12.2 The internal models (class 1) and the system of pictorial organization 

 261

Table 17. The pictorial organization of the Adam and Eve cycle with respect to representations of God and related (extraterrestrial) figures Figure

Positive

Mode of representation Major (God)

Detailed Reduced

Negative

Minor (Michael)   Minor (Messenger)   Major (Satan)

 

Type of scene State/event

Action

(Pii) P11, P41.2

(P2) P9, P10, P41.1, P44, P45.1 P13, P20.1 (right), P31.2, P34.1, P34.2, P36.1, P39.1, P39.2, P45.2 P9, P46 P31.2, P36.1 P20.1 (left), P24, P28, P31.1 P20.2, P36.2

The system of pictorial organization underlies the Creation cycle no less firmly than the other two discussed above (Table 18). By postulating the same opposition between state/ event and action, one may understand in a principled way the significant distinctions between P7.1 (state) and P6/P7.2 (action), despite their overall similarity (section 4.1). First, while God raises his right hand moderately in P7.1, he lifts and extends his right arm with an addressing gesture in the other two pictures.141 In this respect, God’s pose in P7.1 seems closest to that in Pii and P11. In contrast, the gesture in P7.2 corresponds well to that in P2. Moreover, in P6 God points downward with his left hand, and in P7.2 he raises his left hand while grasping a book distanced from himself. In P7.1, however, God assumes a normal posture of holding a book in his left hand. Thus, God’s left hand stays at rest in P7.1, but assumes a deliberate pose in the other two pictures. Second, the bowl is depicted being held by the angel and issuing rays of light in P6/P7.2, whereas a mandorla appears in its usual form instead in P7.1, without movement of light. The grand, pointing gesture of the right hand, the additional posture of the left hand, and the emitting light all give a sense of dynamicity and action to the scenes concerned, while the absence of these dynamic moments represents P7.1 as static and stable. These notable distinctions between P7.1 and P6/P7.2, therefore, may be explained by subsuming the two groups under the opposite categories of state and action, respectively. This identification is corroborated by the parallelism in hand gesture of P7.1 and P6/P7.2 to Pii and P2, respectively. As it turns out, P7.1 is to P6/P7.2 as Pii is to P2. At this point, one may be led to ask: What about the relationship between P7.1 and Pii on the one hand, and P6/P7.2 and P2 on the other? Given that Pii and P2 are of a ubiquitous presence, preexisting and independent of particular cycles, there are two possible analyses. First, P7.1 and Pii, and P6/P7.2 and P2, respectively, are of an even 141 Concerned with an entirely different issue, Hilmo (2004: 88) draws attention to the same distinction between P7.1 and P7.2 with respect to God’s right arm.

262 

 12 The Touronian Foundations of the Old English Genesis: A Synthesis

status in the categories to which they belong. In other words, no additional categorial distinction is in force between P7.1 and Pii, and P6/P7.2 and P2, respectively, as between P11 and Pii or between P10 and P2 in Table 17. Inasmuch as no empirically significant differentiation is feasible between them, the assumption of their homogeneity may be accepted as tenable. Alternatively, one may posit categorial boundaries between the otherwise equal members. That is, instead of two categories, state (Pii and P7.1) versus action (P2 and P6/P7.2), one may postulate a further distinction between detailed and reduced as well. Insofar as substantive differentiation is demonstrably valid that may be plausibly attributed to the categorial distinction proposed, the second analysis must be adopted at the cost of the first. The crucial question is whether P7.1 and P6/P7.2 can be distinguished from Pii and P2, respectively, in such a way as to explain the alleged differences as effects of the categorial distinction between reduced and detailed. A comparison of the figures of God in all these pictures will substantiate the distinction between detailed and reduced. In contrast to Pii and P2, P7.1 and P6/P7.2, respectively, depict God in reduced form in terms of figural size and integrity (section 4.2). In the latter three Creation pictures, God is depicted much smaller and not fully positioned. In P6, God is shown seated, but the throne is absent. In P7.1 and P7.2, the standing figure of God is not wholly represented, with the feet hidden behind the lower dome. Contrary to Pii and P2, no figures of attending seraphim appear in P7.1 and P6/ P7.2. Of no less importance, the frame is minimal in P7.1 and P6/P7.2, a thin arc. In contrast, in Pii and P2, the throne is depicted, and an architectural frame is supplied in Pii, although P2 looks simplified in this respect, yet more substantial, having double lines, than P6/P7.2. Incidentally, the reduced status of the figure of God in P7.1 is confirmed by comparison with P11, the detailed variant belonging to the category of state in the Adam and Eve cycle, particularly with respect to the architectural frame in contrast to the simple geometrical enclosure. One should be reminded that the comparison is motivated by the fact that P7.1 and P11 commonly depict God beholding his creations from heaven. These consistent differences along the dichotomy of detailed and reduced modes of representation therefore may justify postulation of the binary distinction on that basis. Thus, the three pictures affiliated with the Creation cycle, along with the two cycle-independent pictures, are organized by the same system that underlies the other two cycles. Note in passing that, since God is solely responsible for creation, there is no conceptual room for God’s antagonist here. Accordingly, the distinction between positive and negative, and by implication the subdivision into major and minor, are immaterial for this cycle. Table 18. The pictorial organization of the Creation cycle with respect to representations of God  

State

Action

Detailed Reduced

(Pii) P7.1

(P2) P6, P7.2

12.2 The internal models (class 1) and the system of pictorial organization 

 263

Thus, the whole collection of the first twenty-two pictures and their constituent smaller ones in Junius is integrated into the network of images on the basis of the few formal parameters bearing on the ways in which God and other extraterrestrial beings are represented. Since there is nothing of logical necessity in organizing all pictures along the axis of God’s images in the ways determined by the system, specifically on the basis of Pii and P2, at issue here are empirical findings and generalizations, rather than vacuous stipulations posited after the fact. Conceptualized as the deepest, most abstract level of organization, this underlying system provides bases for individual relations, including derivational ones, that are further determined in conjunction with additional conditions and specifications. Accordingly, not all relations prescribed in the system are derivational in nature; derivations constitute a subset of all relations established in the underlying network of images in most abstract terms. Insofar as given images are genealogically linked to each other on the basis of their specific correspondences in form and substance, the relationship mediating them may qualify as derivational in its functionality. Concrete examples may be useful in clarifying this point. As argued above, the Covering group of P34.1, P34.2, P36.1, P39.1, and P39.2 are members of the category (positive, major, reduced; Table 17). This category contains four other members as well, P13, P20.1 (right), P31.2, and P45.2. Common membership in the same category, however, is not equal to their overall derivational identity. The latter four are not covariants of the Covering group, nor are they related to each other in derivational terms: P13 is internally derived from P11; P20.1 (right) is based on an external source, Grandval 5v2 (right); likewise, P31.2 is externally originated from Bamberg 7v2 and in turn served as an internal source of inspiration for P41.2; and P45.2 is a derivative of P45.1 (Table 15). Despite the absence of their connections through derivation, their subsumption under the same category, needless to say, is empirically motivated, as they are all formally founded on the common reductive manner of God’s figure, in comparison with their corresponding images in the opposite categories of detailed representation. Yet P31.2 and P34.2 are not totally unrelated in derivational terms. P34.2 is derived from a common archetype behind the Hortus Deliciarum and the San Paolo fuori le Mura fresco (section 9.2), the symmetrical variant to Bamberg 7v2, which inspired P31.2. In other words, P31.2 and P34.2 may be comparable to cousins in kinship. Distinguished among the set of derivational relations thus delimited are the internal and external derivations, as listed under Table 15 above. The distinction resides in whether a derivational mapping assumed seems to be plausibly traceable to an external source. In the absence of any plausible connectivity to the outer world beyond the confine of Junius 11, the derivation at issue must be regarded as motivated primarily on internal grounds. By contrast, where an external model can be viably postulated for a given picture, it is legitimate to identify it as derivative of that outer source. The categorization as internal is in no small measure a matter of contingency, depending crucially on transmitted material. One of the major empirical findings and generalizations presented here is that the absence of external primary models is coterminous with

264 

 12 The Touronian Foundations of the Old English Genesis: A Synthesis

the absence of corresponding material in the Cotton Genesis family and the Touronian Bibles. As will be fully treated below (section 12.6), Non-Touronian/Non-Cottonian external sources did not serve as primary models for Junius. Now that the system of pictorial organization has been demonstrated to permeate the whole constellation of the first twenty-two pictures in Junius, the next question arising concerns its diachronic basis, that is, its origins and development into the extant synchrony established in Junius. More specifically, where does the system grounded on the specific parameters originate from, and what would have given impetus to its construction in the existing form? As will be detailed below, it is claimed that the pictorial system at issue is, in large measure, derived from the Touronian Bibles as implicitly and opaquely embodied in the Vivian and San Paolo Bibles, through the elaboration and systematization that the Junius artist ingeniously implemented. Thus, far from being created from scratch, the basis of organization was inherently available as raw material in the Touronian models. What the Junius artist accomplished, to his credit, was his penetrating appreciation of latent features as a resource for organization and his masterful building of the system on such a basis. The contrast between state and action and that between detailed and reduced inhere in Vivian and San Paolo in their primitive form. They are already there, still to be elaborated and systematized for full functionality as a basis of overarching organization. The rudimentary distinctions predicated on these two features are manifested, as in Table 19. Table 19. The incipient system of pictorial organization in the Touronian Bibles  

State

Action

Detailed Reduced

San Paolo 1r, 259v, Vivian 329v, 423r San Paolo 117v, Vivian 130v

San Paolo 170v, Vivian 215v, 423r — 

The above three-way contrast is only of latent significance. It is strictly local in its scope of operation and less than rigorous in its patterning: only a small number of pictures are covered, and the contrast appears to be far from a consistent opposition. In short, it remains simply a potential for structuration and will be bound to remain so, unless creative refinement and adjustment are executed on it. Specifically, with respect to the opposition between state and action, the enthroned king (Charles the Bald; Vivian 423r [Pl.  32] and San Paolo 1r) and Christ in Majesty (Vivian 329v [Pl. 30] and San Paolo 259v [Pl. 36]) are contrasted with the acting (dancing) King of David (Vivian 215v and San Paolo 170v). Such postulation of the structural relation involving the distinct personae would ultimately have been grounded on the typological identification of the trio—Charles the Bald, David, and Christ—embodied in both verbal and visual terms by the Vivian Bible as a presentation copy (Kessler 1977: 109,

12.2 The internal models (class 1) and the system of pictorial organization 

 265

129; Kaspersen 1988: 87; Diebold 1990: 85–128; Dutton/Kessler 1997: 71–87).142 One must be reminded in this connection that, in contrast to the corresponding presentation miniature in San Paolo (1r), the Vivian counterpart (423r) is inherently ambivalent between state and action, as it condenses multiple scenes of the ongoing ceremony of the presentation of the Vivian Bible itself to the king (section 2.1). Accordingly, this picture in and of itself embodies the contrast between state and action, thereby affording a common model for the foundational pair of Pii and P2, as detailed in section 2.2. Analogously, the opposition between detailed and reduced is observed among the four examples of God/ Christ in Majesty: Vivian 329v (Pl. 30) and San Paolo 259v (Pl. 36) are evidently detailed in depiction mode, whereas Vivian 130v (Pl.  28) and San Paolo 117v (Pl.  35) may be viewed as reduced in their own different ways, but commonly by virtue of their smaller size. Of particular interest in this regard is the contrast between San Paolo 259v and San Paolo 117v (top register), which seems to be strikingly parallel to the distinction between Pii and P17.1 in terms of composition. Of the possible four-way distinctions, the intersection of action and reduced is empty and unrealized, in the absence of relevant exemplification. Thus, the three-way distinction, less than fully integrated, may be recognized as being in function in the Touronian Bibles. Moreover, the opposition may involve distinct figures, between God and man, as in Vivian 329v and Vivian 423r, thereby rendering the contrast less than full and maximal. In this connection, Vivian 423r (state) and San Paolo 170v (action) may count as an almost optimal contrast, as the figures involved are both human kings, flanked by bodyguards, and situated within the closely corresponding architectural frames. Yet the contrast obtains between the two distinct works. Given that the Junius artist would have had free access to an array of Touronian Bibles, he undoubtedly would have had little difficulty recognizing the contrast in question crossing manuscript boundaries. Building on the germinating pattern of contrast that he took cognizance of in the Touronian models, the Junius artist turned it to virtual perfection by maximization and optimization, and generalized it as a grand design of pictorial organization, culminating in the fully distinctive, four-way opposition system (Tables 16 through 18). He would have composed a series of pictures drawing on this paradigm—used as the internal

142 Charles is addressed as David in the accompanying dedicatory poems (Diebold 1990: 123; Dutton/ Kessler 1997: 81). There are four such poems in the Vivian Bible: Poem I (fols. 1r–2v), Poem VI (fol. 329r), Poem X (fol. 422r), and Poem XI (fol. 422v); Dutton/Kessler 1997: 7–8. Correspondingly, the king is visualized in Vivian 423r with attributes characteristic of David depicted in Vivian 215v, such as the crown, bodyguards, and personified virtues (Diebold 1990: 124; Dutton/Kessler 1997: 81). David in turn is represented as comparable to Christ, as he appears in a mandorla (Vivian 215v) corresponding to Christ in Majesty (Vivian 329v; Diebold 1990: 127), and is described as speaker of Christ’s many mysteries (Poem I. 78; Dutton/Kessler 1997: 68). Finally, Charles is likened to Christ by use of appellatives, such as the splendor populi, lux mundi, and gloria regni, privileged epithets of Christ (Poem  I.  191; Diebold 1990: 121–122; Dutton/Kessler 1997: 84).

266 

 12 The Touronian Foundations of the Old English Genesis: A Synthesis

models—so that, as a guiding principle of composition, the generalization motivated and constrained the artist’s subsequent practices.143

12.3 The external models (classes 2 through 4) As regards external materials in general (i.e., classes 2, 3, and 4; Table 14), almost all pictures are derived through conceptual conversion from the original models that represent different scenes. Exceptions are few: P6/P7 (Creation; but with extensive recomposition), P9 (Creation of Eve), P10 (Admonition/Blessing), P34.2 with several variants (Covering/Hiding; with a wide range of variation), and P46 (Angel Guarding the Gate with a Sword Upraised). The three classes of external sources may be ranked in terms of degrees of conceptual conversion implemented, that is, the extent to which conceptual change occurs between source and target in the course of adaptation, as follows: Scale of conceptual conversion: class 2 > class 3 > class 4 According to the scale, Pii, P2, and P3 (all being class 2) are attended by a maximal degree of conceptual conversion, as can be readily confirmed: the conceptual distance between God and Charles the Bald enthroned or between the Fall of the Rebel Angels and the Conversion of St. Paul is obviously farther removed than that between the Temptation of Eve and the Temptation of Adam or between the Judgment and the Reproval/Denial. In cognitive terms, classes 2, 3, and 4 are based largely on metaphor and/or multidimensional analogy, that is, varying conceptual, compositional, and figural correspondences that the artist would have appreciated as being significantly at play between source and recipient materials. Furthermore, the system and principles of pictorial organization that the artist was committed to—as substantiated in the preceding section—provided an overarching framework for adaptation. In this sense, the mechanism that was in force underlying class 1 was also implicated in the reworking of external sources, although in secondary operation. By bringing to light innermost correspondences otherwise unrecognizable and thus bridging maximal conceptual gaps otherwise unbridgeable, the ways in which the artist reworked extant materials for scenes unique to the Old English/Old Saxon Genesis may reveal his creativity and ingenuity most manifestly, and hence determine the identity of his work most profoundly. Accordingly, of foundational importance is the treatment 143 It may be logically assumed that more than one artist, working successively at different times, might have been involved in the composition of the twenty-two pictures. That is, the system of pictorial organization might have been inherited and enriched through transmission and learning. Since this possibility is conceptually more complex by containing extra hypotheses totally lacking empirical support, one may well leave it out of consideration.

12.3 The external models (classes 2 through 4) 

 267

of the Fall of the Rebel Angels cycle (P2, P3), along with the frontispiece (Pii), which is also extraneous to the Book of Genesis. Because the Fall of the Rebel Angels cycle is totally foreign to the Book of Genesis, it would be impossible to find ready-made models within the Cotton Genesis tradition. As a consequence, sources must be duly sought outside the Cotton Genesis. That is, entirely based on Touronian/Non-Cottonian materials, only these pictures fall under class 2 according to the above classification of sources (Table 14). Among the Touronian Bibles, the Vivian (and to a lesser extent Pre-San-Paolo) models figure exclusively. The noninvolvement of the Bamberg and Grandval Bibles is simply due to the absence of relevant materials there, as neither the Jerome cycle nor that of St. Paul is materialized in them. A presentation miniature, too, is absent in Bamberg and Grandval. Thus, the monopoly of Vivian should be only apparent. By contrast, since the cycle of Adam and Eve is shared in its essentials between the original and its vernacular reworking, it would be natural to keep using the traditional materials by supplementing alterations and adjustments as deemed necessary. Therefore, it may follow as a matter of course that the whole operations of adaptation for the Adam and Eve cycle were implemented on materials drawn from the Touronian/ Cottonian sources (class 3) most readily available. The Touronian/Cottonian models thus chosen are not restricted to Vivian; rather, Bamberg is featured prominently. The lack of bias toward a particular Touronian Bible, notably Vivian, may confirm the above inference that the privilege of Vivian in class 2 is only apparent due to the ample availability of source materials. It further shows that the Junius artist would have had equal access to all Touronian Bibles, including the Touronian antecedent of the San Paolo Bible. In purely logical terms, however, the Adam and Eve cycle could have been composed on the Non-Touronian/Cottonian (class 4) basis, that is, by direct and exclusive references to the Cotton Genesis materials in distinction from their specific embodiments in the Touronian Bibles. As a look at Table 15 (under class 3) indicates, there are two instances in which such Non-Touronian/Cottonian materials—San Marco—seem to have been invoked as primary models: P24 and P31.1. Yet in both, the Touronian Bibles were consulted concurrently. Since no distinction can be made between the Touronian and Non-Touronian models in compositional terms, however, it is impossible to determine which group, Touronian/Cottonian or Non-Touronian/Cottonian, would have prevailed as source material. Another noteworthy group subsumed under the Adam and Eve cycle concerns P31.2 and P41.1/P44. In it, the Touronian Bibles are identified as primary models, whereas the Non-Touronian/Cottonian corresponding images, specifically the San Marco reflexes, are definitely distinct and do not figure at all. Insofar as the Non-Touronian/Cottonian instances are taken to be faithful descendants of the Cotton Genesis archetype, the selection of the Touronian manifestations, presumably more deviant from the archetypes, may be construed as convincing evidence for the primacy of the Touronian Bibles as sources for Junius to the sacrifice of the Non-Touronian/Cottonian examples. P31.2 (and derivatively P41.2 as well; see section 10.2 above) may be viewed as a case in point here. While Bamberg was chosen as a primary model for P31.2 (and P41.2 through the

268 

 12 The Touronian Foundations of the Old English Genesis: A Synthesis

intermediary of P31.2), the corresponding San Marco image was not adopted. The chief difference between the two candidates is the crouching (Bamberg) versus the standing (San Marco) pose of Adam and Eve. That the standing posture seems more likely to have been closer to the archetype is suggested by the symmetrical variants of the Covering, as exemplified in San Paolo and the Hortus Deliciarum, which display standing figures as well. In this light, it may be inferred that the crouching attitude would have been a secondary development in Bamberg. Despite the deviation from the archetypal composition, however, Bamberg was used as a model for Junius. Also to be recalled in this regard is that an adjacent figure was incorporated into the Junius picture, the configuration that is unique to Bamberg. These considerations may lead one to conclude that, where options were available, the Touronian/Cottonian materials were chosen over the Non-Touronian/Cottonian counterparts, even when the latter are more heavily based on a Cotton Genesis copy. Much the same is observed in the competitive treatment of secondary models between Touronian/Cottonian and Non-Touronian/Cottonian sources. Underlying P9 is a complex of scenes in Vivian 10v1—the Creation of Adam, the Observing Angel, and the Drawing of a Rib from Adam. In addition, the mirror image of the succession of the Drawing of a Rib and the Introduction of Eve to Adam in Bamberg 7v2 was invoked as a supplementary model to this grand design. Of particular significance in this connection, Bamberg 7v2 would have been chosen at the expense of an alternative model, the mirror image of the Creation of Eve consisting of the Drawing of a Rib and the Formation of Eve in the San Marco mosaics (Pl. 43), the sequence that matches P9 perfectly in conceptual terms. Despite such complete agreement, the San Marco image was not adopted. This is all the more remarkable, because the San Marco picture is closer to the Vivian model in figural details as well, such as God holding a cross-staff, and God and Eve/Adam standing at the same ground level. Thus, the conceptual and compositional identities notwithstanding, the less similar source, Bamberg 7v2, was prioritized over the otherwise perfect candidate, San Marco. This is accordingly another empirical confirmation of the privilege accorded to the Touronian/Cottonian sources over the Non-Touronian/ Cottonian counterparts by the Junius artist seeking inspiration. If there were, on the contrary, grounds for assuming that the Non-Touronian/Cottonian images at issue display lesser degrees of conformity to their original, the above reasoning could not be pursued convincingly. It may be inferred that the apparent mismatch between the Non-Touronian/Cottonian and Junius images in contrast to a significant parallelism between the Touronian/Cottonian and Junius ones would have been due to transformations introduced into the Non-Touronian/Cottonian sources subsequent to the production of the Junius images; in other words, at the time of the composition of Junius, the Non-Touronian/Cottonian sources may have been no different from the Touronian/Cottonian realizations. A case in point falling under this category is P41.1/ P44. As argued at length in section 12.4 below, the symmetrical arrangement of Adam and Eve characteristic of the San Marco representation would have been an innovation by the San Marco mosaicist, inspired by P41.1/P44. Therefore, the resulting deviation

12.3 The external models (classes 2 through 4) 

 269

from the Junius scene and its primary models, Bamberg and Grandval, would have been a secondary phenomenon. Thus, class 3 (Touronian/Cottonian) is by far the most extensively used among the three external sources for image creation (Table 15). Of further significance, while ultimately affiliated with the Cotton Genesis family, this class is primarily and specifically based on the Touronian manifestations of the family. Correspondingly, by virtue of the limited materials in the Touronian Bibles relating to the Book of Genesis (i.e., the Adam and Eve cycle represented in the Genesis frontispieces), the adaptation of the Touronian/Cottonian models is not confined to the images of the strictly comparable motifs, such as the Admonition/Blessing derived from Grandval, remarked at the outset of this section. Actually, this is a rare case not only because of the immediate conceptual correspondence between source and target, but also because of the limitation to a single source (see the following paragraph). The reuse of related—in the sense of belonging to the same Adam and Eve cycle—but distinct—in the sense of not visualizing the same scene types—images in the Touronian Bibles for other scenes is more widespread than the strict adherence to the original motifs. This is hardly surprising, given that the Old English/Old Saxon Genesis deviates so markedly from the Book of Genesis in its narrative contents: the Fall of the Rebel Angels is a novel story in the context of Genesis, and the Temptation of Adam and Eve, while common overall, is attended by distinct narrative components, such as Satan’s messenger disguised as an angel tempting Adam first. Regardless of the exact categorization of external models chosen (classes 2 to 4), the adaptive methods and techniques executed are susceptible to general accounts. Moreover, basically the same observations may apply to the auxiliary as well as the primary models, as exemplified below. As mentioned in the preceding paragraph, one only rarely observes a strict one-to-one correspondence both in form and sense between source and target images, a borrowing that would be close to a mechanical copying. Usually, the Junius artist brought multiple sources together to compose a single picture. Put another way, the resultant Junius images are composites based on cognate materials drawn from more than one source. The multiplicity of sources may fall into two types. First, more than one picture may be adopted from a single work. Second, more than one work may be consulted, whether limited to the same single episode or involving multiple ones. The grand layout for the Creation of Eve in P9 would have been inspired by the succession of the Animation of Adam with a Watching Angel and the Drawing of a Rib from Adam, the two separate scenes that occupy the upper left half of the first register in Vivian 10v1. The derivation of P31.2, Adam and Eve Praying and Lamenting, from Bamberg 7v2 (far right, Covering), in conjunction with the figure of Eve Taking and Giving the Fruit in Bamberg 7v2 (second from the right), may count as a further intricate example of the first type. The crucial figures underlying the trio of Adam, Eve, and Satan’s messenger in P31.2 are distributed in two adjacent scenes across the panel border in the original. Moreover, while appearing in the scene next to the Covering in Bamberg, the figure of Eve that

270 

 12 The Touronian Foundations of the Old English Genesis: A Synthesis

served as a model for the devil in P31.2 is not found in contiguity to the covering figures of Adam and Eve in Bamberg, but intervened by another figure, which is disregarded as immaterial for the composition of P31.2. The other variant of the second type, a synthesis of corresponding scenes found in more than one work is exemplified in its simplest form by the double depiction of the snake in the Judgment (P41.1), in which both the standing and creeping figures of the snake appear in their dynamic concatenation. This double figuration naturally may be explained as a compositionally simple case of an additive operation whereby the significantly varying elements in the otherwise almost identical materials—the creeping snake (Bamberg) and the standing snake (Grandval)—are each incorporated by compounding into P41.1 as temporally distinct phases of the same entity. Similar, but going beyond small local constituents to cover the whole pictorial domain, is P24. Depicting the Temptation of Adam, the image draws on two corresponding scenes in two different works: the Animation of Adam in the San Paolo Bible and the San Marco mosaic. Each of the sources posited contributed in its own indispensable way to the construction of a new image in Junius. By contrast, representing the conflated scene of the Temptation of Eve and the Fall of Man by the specific configuring of the snake, Eve, and Adam, both Bamberg 7v2 (middle) and Vivian 10v2 (left) resist differentiation from each other, and therefore both must be regarded as primary models for P28 in the same capacity (section 8.2). Since these two images are essentially indistinguishable and since the Junius artist did not consistently favor one work over the other in his selection of materials elsewhere, it should be legitimate to acknowledge both as equally responsible for the recomposition in Junius. A similar composite image creation would have been involved in nonprimary models as well. While derived primarily on internal grounds, P11 received inspiration from many other auxiliary sources (section 6.2). As may be recalled, the whole picture can be divided into two constituents, God enclosed in an architectural frame and the landscape of paradise, particularly the representation of animals and birds. Each constituent draws on multiple sources taken from more than one work. While the animal scene is traced back to the two cognate representations, the Naming of the Animals in Bamberg 7v1 and Millstatt 9r, the figure of God framed in a hexagonal architectural structure would have resulted from a more complex derivational history involving a wider range of heterogeneous sources defying precise identification. Particularly at stake are an array of Christ in Majesty as in Vivian 329v and San Paolo 259v, and the city walls and buildings with the gate on the viewer’s side, as in Vivian 386v2 (right) and San Paolo 3v2 (left). While a synthesis of multiple source images normally results in a unification of the corresponding figures involved, it may sometimes happen that the original figures each find a distinct manifestation in a resultant picture. This exceptional treatment may be observed in the representation of Lucifer in P3.1 (section 3.2). The central large figure of Lucifer is partly derived from the beckoning pose of Jerome in Vivian 3v1 (left) and partly from the two astonished attendants in Vivian 386v1 (center)/San Paolo 310v1

12.3 The external models (classes 2 through 4) 

 271

(center), whereas the smaller one to the upper left of Lucifer in the center, about to enter the palace, is indebted to the San Paolo 3v1 (left) counterpart just boarding the ship, as well as to the Vivian 386v1 (right) figure of Saul entering Damascus. Thus, rather than being integrated into a single representation, the two different figures of the same man (St. Jerome) in the two works receive visualization as two distinct figures of the same being—Lucifer—rather than unified into one. This split individuation of the same identity may have been largely facilitated by the multiple appearance of the same man in the model panels, that is, Vivian 3v1, San Paolo 3v1, Vivian 386v1, and San Paolo 310v1, and also by the contributions of other characters—Saul and his attendants—to the same figuration. In this light, the apparently peculiar adaptation proves to be appreciable as a reasonable method of visualization fairly faithful to the models at a higher level of pictorial organization. The adaptive uses of inherited materials considered so far presuppose the correlation of variation in images to differing scenes, whether a given conventional association between image and sense was reproduced or (more frequently) transformed in the course of adaptation. Accordingly, where two distinct images are recognized, they are interpreted as likely visualizing two different scenes. Among the stock of images inherited in the tradition, however, are those that lack their own signification in correlation to their formal variation on the surface. The differences observed in such cases are purely matters of outer variation without corresponding inner conceptual distinction. These nondistinctive images may therefore be identified as mutual variants lacking autonomous status as independent images, and are subsumed under a common image type that is manifested in such variable forms. At times, however, the Junius artist challenged such a conventional lack of semantic differentiation between purely formal variants and established a conceptual distinction anew where there had been none before. A case in point is the splitting in Junius of the conventionally single scene type of the Covering/Hiding into the Praying (P31.2) and the Covering (P34.2), as discussed in depth in section 9.2 above. Behind this image bifurcation is the traditional variation on the Covering/Hiding with respect to the orientation of Adam and Eve: on the one hand, Adam and Eve may face the same direction (e.g., Bamberg 7v2, San Marco Pl. 47), and on the other they may be arranged facing each other (e.g., San Paolo 8v2, Hortus Deliciarum 17v). Traditionally, each of the Cotton Genesis members selected one of the two variants without making a conceptual distinction concomitantly. Since both variants were not adopted in the same work, no possibility arose for differentiating meanings according to formal variation. The Junius artist, however, revolutionized the conventional treatment by introducing both variants, conferring independent status to each, and accordingly individuating two separate images loaded with two distinct meanings. Forging a correlative conceptual differentiation on the basis of a purely formal variation current in traditional practice would have necessitated a familiarity with such a wide variety of source materials as to ascertain the existence of variable images inscribed with no corresponding semantic distinction. A simple, casual contact with a model at

272 

 12 The Touronian Foundations of the Old English Genesis: A Synthesis

hand doubtless would have been insufficient for such innovative reorganization and reinterpretation. The exact ways of treating source materials are at times very sophisticated in operation. In constructing the Creation of Eve (P9), the artist drew on a supplementary source, Bamberg 7v1, in addition to Vivian 10v2, which lays out the grand scheme including a depiction of an observing angel (see above). What is remarkable in this case is the use of a horizontally reversed image from the Bamberg material, whereby the earlier scene of the Drawing a Rib from Adam is repositioned on the right side. Much the same technique, but compounded with a further elaboration, is exercised for the Judgment (P41.1/ P44): as far as the Reproval/Denial—with respect to the configuration of God, Eve, and the snake—is concerned, a mirror image of Bamberg 7v3 and Grandval 5v3 is reconceptualized as the Judgment of each (P41.1 and P44 left), whereas the original orientation is maintained for God and Adam yet reinterpreted in parallel fashion (P44 right). Thus, the artist had access to not only actualized representations, but also to virtual ones in the mind obtained by a series of cognitive operations such as resegmentation, rearrangement, relocation, reorientation, and reduplication. By far the most complex composition is that of the two figures of Lucifer in P3.1. Drawing on the two Touronian Bibles—Vivian and the Touronian predecessor of San Paolo—and making eclectic use of the two distinct scenes each represented in both works, the artist arrived at the two synthetic representations of Lucifer that he arranged side by side in P3.1, as mentioned above from a slightly different perspective. On the one hand, the central large figure of Lucifer inviting his followers to his palace was created, through reconfiguration, from Saul’s two astounded attendants in Vivian 386v1 and San Paolo 310v1 (two cognate scenes in the Conversion of St. Paul) and the beckoning figure of Jerome in Vivian 3v1. On the other hand, the figure of Saul facing the high priests (San Paolo 310v1) and being ushered into the City of Damascus (Vivian 386v1, right), and that of Jerome about to board the ship (San Paolo 3v1) are transformed, through integration, into the other figure of Lucifer about to enter the new palace. The artist’s ingenuity, however, was not exhausted at the directly visible level of adaptive figuration, based as it was on keen observation and imaginative use of source materials. Transcending surface examination, the artist masterfully appropriated a representation technique practiced in Vivian 386v1, in which the fallen figure of Saul instantiates a double role by absorbing a standing self at the moment of attack. Once familiarized with the technique, the artist quite productively applied to the central figure of Lucifer in P3.1 a comparable conflation of temporally distinguished double figures into a unitary representation by compounding attributes originated from the distinct two: the rebel leader is pointing to the new palace from outside, on one side, and receiving crowns from his subjects inside the palace, on the other. The Junius artist’s masterful appropriation of a marked technique of figural synthesis may reconfirm the firm basis on and the profound indebtedness to the Touronian tradition of visualization in his practice. The artist’s engagement with the Touronian Bibles in the production of Junius reveals the extensiveness of materials covered and the depth

12.3 The external models (classes 2 through 4) 

 273

of insight into and understanding of the underlying principles of pictorial organization in the Touronian models. A casual, superficial acquaintance with an isolated instance of the Touronian Bibles most unlikely would have made the admirable accomplishments possible. Therefore, it may be plausibly inferred that the artist was likely a practicing member of the School of Tours. The Touronian/Cottonian models for the Adam and Eve cycle in Junius are all progressing from left to right in a register. This consistent rightward narrative development, however, is usually reversed within a single picture in Junius that accommodates a temporal sequence of separate scenes in its own pictorial domain. Specifically at issue are P9, P20, P28, and P41.1, which all display a right-to-left succession of events/ states, obviously at odds with the original sequencing from left to right in the Touronian models. By contrast, there is only a single instance of left-to-right orientation, P44.144 As substantiated in section 10.1, however, this exceptional directionality may have been motivated by another compositional factor, namely, to establish optimal correspondence and contrast with P28. Given the principled reason for the sole exception, it may be generalized that in Junius the original left-to-right narrative orientation is consistently reversed to the right-to-left counterpart. The fully complementary distribution thus observed between Junius and the Touronian sources displays a statistically significant difference, as a Fisher test performed on the contingency table of 4/0 versus 0/4 gives a p-value of 0.029. Thus, the divergence in pictorial directionality is a genuine empirical issue requiring principled explanation. Why did the Junius artist decide to reverse the direction of progression? Was there any model for this innovation? Rather than addressing these questions directly, one may take a roundabout way to explore what at first sight appears to be a totally unrelated problem: the relative position of God in a picture, more specifically, where God is positioned on the horizontal axis in a relevant pictorial domain, on the right side or the left, and in which direction God faces, to the right or to the left. Thus, one may be interested in the location and orientation of God in terms of right/left polarities in the Adam and Eve cycle in Junius and the Touronian Bibles. Accordingly excluded from consideration are representations in which God is situated in the center facing the viewer. As will be shown presently, the two parameters fully overlap: the figure of God standing at the right faces left, while the one standing at the left faces right. Thus, the two parameters are redundant and reducible to a single variable. This common binary variable—ranging between “standing right/ facing left” and “standing left/facing right”—is distributed in the Adam and Eve cycle in the five works under consideration, as indicated in Table 20.

144 All other pictures subsumed under class 3 are irrelevant for the present concern, as they embody no horizontal progression.

274 

 12 The Touronian Foundations of the Old English Genesis: A Synthesis

Table 20. Distribution of figures of God in the Adam and Eve cycle, according to their horizontal locality/ directionality and in terms of token frequencies  

Standing right/facing left

Standing left/facing right

Progression

Junius Bamberg Grandval Vivian San Paolo

3  3  0  1  5 

0  2  5  3  1 

right to left left to right left to right left to right left to right

By way of offering concrete descriptions for each of the works, the right-standing, left-oriented figures of God appear in P10, P41.1, and P45.1 in Junius. In addition, there are two other instances (P9 and P44) in which God is doubly depicted within a single picture in such a way that in one of the constituent pictures God appears left and faces right in symmetry to the other pair standing at the right and oriented left. Such components, however, lack autonomy as self-contained pictures, as they are fully integrated with each other compositionally without explicit boundaries for motivating their ­separate conceptualization. Therefore, it may be generalized that there are no examples in Junius in which God, uniquely depicted, stands at the left and faces right. Further, reformulated with an exclusive reference to face orientation, this generalization holds not only for the Adam and Eve cycle as has just been observed, but also for the Fall of the Rebel Angels cycle (P2, P3.2b, and P16.1). With respect to Bamberg, the right-standing and the left-standing figures are represented with similar frequency. Of further interest, God appears at the right and faces left in the first three pictures, while he stands at the left and faces right in the remaining two. In Vivian, God is shown standing left and o ­ riented rightward in the first three scenes (the converse of the Bamberg pattern), but only in the Reproval is he positioned at the right and facing left. Grandval consistently depicts God standing left and facing right in all of the five relevant scenes. Finally, San Paolo exhibits the strongest preference—five pictures in all—for the right-standing/left-facing figure of God among the Touronian Bibles, with a single exception (Introduction of Eve to Adam). Given the above varying distributions among the five works, one may postulate the following scale of preferences for the right-standing/left-facing figure of God: Preferences of the figure of God for standing right/facing left:  Junius > San Paolo > Bamberg > Vivian > Grandval Junius and Grandval stand at polar opposites: while Junius maximizes the right-standing/left-facing figure of God, Grandval categorically excludes it in favor of the left-standing/right-facing counterpart. In statistical terms, their distribution patterns differ significantly (p = 0.018, by Fisher test). While Junius and Grandval are categorically incompatible, each maximizing its own privilege at the expense of the other’s, they are comparable in their strict conformity to a single orientation, to the exclusion of the

12.3 The external models (classes 2 through 4) 

 275

other three Bibles. At this point, the issue of narrative progression assumes relevance. Only in Junius and Grandval is this parameter—right to left versus left to right—perfectly matched with God’s position/orientation. The right-to-left sequencing is isomorphic to the left facing in Junius, and conversely the left-to-right procession corresponds optimally with the rightward orientation in Grandval. To the extent that the other three members of the Touronian Bibles compromise by allowing for the left-facing posture in the overarching left-to-right progression, they deviate from optimal harmony between the two structural parameters. One may conjecture that the Junius artist would have striven to achieve an optimal correspondence between these two parameters, as with Grandval. Given the varying deviations actualized from the optimal state, however, there seems to be nothing irrational and malfunctional about allowing for a disharmony between the two parametric specifications. Nonetheless, from a formal–structural point of view, the perfect match no doubt would have been valued as an asset in terms of overall pictorial organization. Thus, the Junius artist’s concern with attaining formal optimality seems to be reasonable and understandable enough, particularly in view of the keen awareness he displays elsewhere in matters of structuration and systematization. A question remains unanswered: Why did the artist choose to place God at the right and orient him leftward, rather than the opposite alignment as in Grandval, one of his major models? Why did he venture to depart here from the overarching principle of organization that is in force in his source of inspiration? While to fully recover the artist’s underlying motivations for the innovation in this respect may be a matter of anybody’s guess, it hardly seems wide of the mark to conjecture that P2 played a determining role in his subsequent execution. In it, contrary to the central position occupied by the left-facing king in the model (Vivian 423r), the corresponding figure of God is placed at the right margin, with the left orientation kept intact. This shift in positioning was prompted by the arrangement of Lucifer in the center, because he is featured as a key figure in this context, particularly in relation to the immediately following complex of pictures (P3), the Fall of the Rebel Angels (section 2.1). It should be noted that the rearrangement of God to the right periphery constituted a radical departure from the Vivian model; rather than a consequence of mechanical copying, this was an innovation of the artist’s own design. This deliberate repositioning of a left-facing God at the right, then, may have assumed a binding force and served as a leitmotif for subsequent pictures. As fully demonstrated in the preceding chapters, the Junius pictures were shaped as substantially by the prior images of its own making as by the Touronian predecessors. Once the impetus had been initiated, a growing accumulation of earlier pictures embodying repeatedly the framing condition at issue would have acquired increasing momentum for regularization so much so that the exceptionless generalization came to be established, controlling not only the pictures derived from the Touronian models (classes 2 and 3; Table 15), but also those based on internal sources (class 1). The Junius artist’s striving for perfect harmony between narrative progression and God’s location/orientation presupposes that he had a keen perception of the two

276 

 12 The Touronian Foundations of the Old English Genesis: A Synthesis

parameters in function and the varying degrees of their correlation in the models, and that he fully appreciated the optimality achieved in Grandval. This profound appreciation implies that the artist would have been thoroughly familiar with the models, going far beyond casual observations on the surface and delving deep into the intricacies of structural organization. In other words, the artist had unqualified access to the working knowledge system that the Tours School artisans shared (see further section 12.8).

12.4 Feeding the Cotton Genesis family: The innovation of P41/P44 The vast majority of Junius pictures that originated from the Touronian Bibles and/or the Cotton Genesis did not demonstrably influence other members of the Cotton Genesis family, as whatever adaptations and rearrangements introduced in Junius were strictly local and internal in their scope. In this respect, the treatment of the Judgment (P41.1/ P44) is entirely different, as it goes beyond the confine of the Junius manuscript in itself to have a significant bearing on the genealogy of the Cotton Genesis family at large. As argued at length in section 10.2, the depiction of this scene is an innovation in Junius through a well-motivated conversion of the Reproval/Denial, concomitant with manifold modifications made. No less significantly, there is reason to believe that this innovated scene was adopted and adapted in the San Marco mosaics, which reconfigured the borrowed scene largely in symmetrical fashion. These claims—that the Judgment scene appeared on its own for the first time in Junius, and that the San Marco representation was inspired by that of Junius, rather than a common descent from the Cotton Genesis archetype—have significant implications that deserve further exploration, in particular when they seem to utterly contradict Weitzmann and Kessler’s reconstruction of the Judgment scene in the Cotton Genesis original (Weitzmann 1984: 107, 115, 141–142; Weitzmann/Kessler 1986: 18–19, 132). According to their account, furthermore, the original scene of the Judgment was subject to recomposition at San Marco in such a way that the three figures of God, Adam, and Eve were rearranged symmetrically by the mosaicist, who intended to make of the Judgment of Adam and Eve a parallel to the Last Judgment. While this hieratic account for the cognitive basis of symmetricalization is indeed credible, what is debatable is their thesis that the Judgment scene existed on its own in the Cotton Genesis archetype. The prime evidence for Weitzmann and Kessler’s claim seems to be the alleged Judgment scene in the Histoire Universelle en Prose (Vienna, Österreichische Nationalbibliothek, Cod. 2576), fol. 4r (left). This is a fourteenth to fifteenth century book that Weitzmann and Kessler (1986: 20–21) categorize as a member of the primary group of the Cotton Genesis, along with the San Marco mosaics and another minor manuscript (Vienna, Österreichische Nationalbibliothek, Cod. Theol. gr. 7; Weitzmann/Kessler 1986: 18–21). While Koshi (1973: 12) read this scene (Histoire Universelle 4r) as God’s Curse (= Judgment), he admitted compositional resemblances to the Hiding (Pl. 48) and the Denial (Pl. 49) at San Marco. The picture should rather be identified as a conflation of the

12.4 Feeding the Cotton Genesis family: The innovation of P41/P44 

 277

Covering/Hiding, the Reproval, and the Denial (the Judgment had yet to be categorized at this stage) in the absence of comparable depictions between which to distinguish. Thus, it is not legitimate on synchronic–structural grounds to posit an autonomous Judgment in the Histoire Universelle, for which the distinction from the Reproval or the Denial or the Reproval/Denial must be established beyond reasonable doubt (section 10.2). As it is, however, the Histoire Universelle does not meet this structural requirement. To be sure, the right image in 3v could be characterized as a Reproval scene, yet the representation of God exclusively with a hand from above in this context is unexemplified elsewhere in the Cotton Genesis family (Koshi 1973: 11). The depiction of Adam to the exclusion of Eve is another peculiarity, given that God’s Reproval involves Adam and Eve as a couple (section 10.2). More substantively, the Reproval presupposes the Covering/Hiding, but this image of Adam apparently being reproved by God does not contain a covering act. Because of these clustering aberrancies, the identification of this unique scene as the Reproval in the Cotton Genesis context is implausible. At the very least, the occurrence of this Non-Cotton style image (cf. Koshi 1973: 11) immediately before the scene under consideration may call into question the validity of viewing the latter as a reliable reflection of an archetypal scene—the Judgment—in the Cotton Genesis. Furthermore, even if one gave the benefit of doubt to this criterion of differentiation from the Reproval, the Histoire Universelle image in question does not appear to qualify as a faithful reflex of the original representation of the Judgment. In addition to the unusual depiction of the rebuking God in this context just mentioned, the snake is not represented in its normal form but in the shape of a dragon, which further betrays this picture’s deviation from the exemplar (cf. Weitzmann/Kessler 1986: 21). Invoking a parallel from Hildesheim as Weitzmann and Kessler (1986: 21n58) do to alleviate the oddity involved only emphasizes the Histoire Universelle’s markedness in this respect. In light of the concurrent anomalies in the alleged Judgment and the immediately preceding picture—the presupposition for its definition as the Judgment—in the Histoire Universelle, it would be methodologically inappropriate to reconstruct a corresponding scene of the Judgment in the Cotton Genesis original, principally based on such a unique reflex far removed in time from the archetype. After all, it is just as likely that the alleged Judgment scene of the Histoire Universelle in itself—if so determined at all—would have been a relatively late innovation in the Cotton tradition, along with its related figures, God’s hand and the dragon. In this connection, it may be interesting to offer a substantive argument against Weitzmann and Kessler’s reconstruction of the Judgment proper in the original Cotton Genesis, with respect to the derivation of the corresponding image at San Marco (Pl. 50). As mentioned above, Weitzmann and Kessler derive this picture from something closest to Histoire Universelle 4r by invoking the mosaicist’s intention to recompose the Judgment of Adam and Eve as parallel to the Last Judgment. Although the underlying motive for symmetrical reconfiguration commands credibility, the alleged trajectory of derivation is questionable. Specifically, there is too much discrepancy to be bridged between San Marco and the Histoire Universelle in their respective representations, apart from

278 

 12 The Touronian Foundations of the Old English Genesis: A Synthesis

the figuration of the snake and the symmetrical arrangement of the keeling couple. That is, God faces leftward toward the snake at San Marco, while he turns right toward Adam in the Histoire Universelle. Following Weitzmann and Kessler’s reasoning on parallelism with the Last Judgment, God’s address to Adam and Eve would be of paramount importance and hence deserve to be retained by all means at the expense of other components of the picture. The actual treatment to the contrary, then, defies explanation. By contrast, one may readily provide a principled account of these features characteristic of San Marco by postulating P41/P44 in Junius—particularly the superimposition of P41.1 and P41.2—as the model for the San Marco adaptation, as demonstrated in section 10.2 above. The conversion of the Denial of Blame to the Judgment is a Junius innovation. That it was subsequently adopted and adapted at San Marco may testify that the innovation was fed back on the Cotton tradition, from which San Marco drew inspiration anew. This demonstrates that Junius (its Touronian/Old Saxon antecedent, to be more precise; section 12.8) was a living member of the Touronian and Cottonian group, interacting with other members. It was not only fed by the tradition, but also fed back to it by enriching the common stock of images available for future use.145

12.5 The unused material in the Adam and Eve cycle in the Touronian Bibles The Adam and Eve cycle in Junius is largely derived from the Touronian Bibles (the external models, class 3; Table 15). It may be illuminating to set a standpoint on the Touronian models and to examine how much of the frontispieces to Genesis was put to use for illustration in Junius. As it turns out, almost all of the Adam and Eve cycle materials available in the Touronian Bibles were reflected in the extant pictures in Junius, except for the following few: the Creation of Adam (Bamberg 7v1, left; Grandval 5v1, left; San Paolo 8v1, left); the Labor (Bamberg 7v4, left/right; Grandval 5v4, right; Vivian 10v3, right; San Paolo 8v3, right); and the Burial of Abel (Bamberg 7v4, middle). The absence of the Creation of Adam in Junius should be only apparent, as the corresponding pages are lost in the extant manuscript (section 5.2). It is unlikely that a scene of such prime importance would have escaped visualization. The Labor scene was obliterated, made redundant by the resetting of the Expulsion (corresponding to the couple driven out by the angel in the Touronian Bibles) in the field and its resultant conflation with and substitution for the Labor. Thus, the omission of the Labor scene would have been caused

145 Another possible instance of San Marco’s indebtedness to Junius is the Temptation of Eve (Pl. 45). As speculated in note 110, section 7.2, the San Marco scene may actually have been based on the upper half of P20 through erasure of Eve from the right side, in accordance with their purely formal derivational relationship elucidated in section 7.2 above.

12.5 The unused material in the Adam and Eve cycle in the Touronian Bibles 

 279

by its assimilation to the recomposition of the Expulsion. This then leaves the Burial of Abel in need of specific discussion. P49, the Story of Cain and Abel, has no correspondence in the Touronian Bibles, as their Book of Genesis contains only pictures for the Adam and Eve cycle in the frontispiece, with the exception of the Burial of Abel in Bamberg 7v4 (middle), and possibly the Slaughter of Abel in Grandval 5v4 (“Beneath the right tip of the garland that forms Eve’s bower” Kessler 1977: 24 and fig. 44). Without a direct textual basis in the Bible, this burial scene is unique, unparalleled elsewhere in the Touronian group and beyond in the Cotton Genesis family at large (Kessler 1977: 25; Klein 1984: 100). Based on the foregoing examinations, one may generalize that the Junius artist consulted the Touronian Bibles first, and only thereafter the Cotton Genesis (section 12.3). Accordingly, whatever motivations would have driven the Bamberg painter to insert the Burial scene in the Adam and Eve cycle—apparently out of context—the Junius artist would have been expected to use the extant image of the Burial of Abel for his work, in line with his treatment of other pictures in the frontispiece to Genesis. All other materials in the frontispiece to the Book of Genesis in the Touronian Bibles were used to good effect, with the exception of the Burial, as shown above. Therefore, the artist would have been expected to adopt the only remaining picture in Bamberg to (re)arrange it at the bottom of P49, immediately following the Slaughter of Abel (see the following paragraph). Actually, however, he did not do so. Instead, he represented a different scene: Abel is depicted appearing out of the earth in its upper body, with an orant gesture toward God, who is standing on the ground. Also being unique in itself, this picture is unlikely to have had a corresponding image in the Cotton Genesis archetype (cf. Weitzmann/Kessler 1986). A question then arises: Why did the artist abandon his strategy of visualization that he had consistently implemented so far, that is, the utmost prioritization of the Touronian images? Specifically, why did he dismiss the readily available image in Bamberg, in favor of constructing a novel one? Before moving forward any further, a brief look at the Slaughter of Abel in P49 (Pl. 71) may be essential. Despite Broderick’s (1978: 256) remark on its resemblance to a corresponding scene in the Roda Bible (fol. 6r), this image seems more strikingly similar to that of San Marco (Pl. 72), especially in regard to the figuration of Abel: while mirror imaged, Abel is depicted lying 45 degrees off the ground, looking up at Cain, and raising one of his arms for protection. Accordingly, it may be assumed that the scene is derived from the Cotton Genesis family, but not from the Touronian Bibles, which do not contain a corresponding image. At this point, the appearance of the Burial of Abel in Bamberg just mentioned, and its absence in Junius, take on great significance, given the exclusive contributions of Bamberg to Junius that are independently attested (the Naming of the Animals for P11/P13; the Hiding for P31.2/P41.2; the Expulsion for P46.2). Furthermore, while the Burial and the Slaughter are not the same scenes, their close relation seems indisputable. Since the Burial presupposes the Slaughter, the exclusive appearance of the former in Bamberg may be viewed as a kind of conflation. Noting that the two scenes are quasi-identical if not perfectly coterminous, one finds here four possible options

280 

 12 The Touronian Foundations of the Old English Genesis: A Synthesis

to choose from in visualization: (i) choose San Marco; (ii) choose Bamberg; (iii) choose both; and (iv) choose neither. Option (i) was actually selected in Junius, as well as at San Marco. Option (ii) gave rise to Bamberg. Option (iii)—selection of both scenes—would seem most reasonable, given the complementarity of the two pictures, strictly speaking. Actually, however, this never materialized. Finally, option (iv) was followed by the Touronian Bibles—except for the Bamberg Bible—in which there is not a single picture affiliated with the story of Cain and Abel. Against such background, the unexpected absence of the Burial scene in Junius may be reformulated more rigorously as the choice of the Non-Touronian/Cottonian (class 4) material over the Touronian/Non-Cottonian (class 2) material, going counter to the well-motivated generalization—the Touronian first principle, as it were. One of my hypotheses, which has its empirical adequacy increased steadily through the advancement of the foregoing investigation, is that exclusively the first twenty-two pictures, up to the end of the Adam and Eve cycle, had access to the Touronian Bibles in their composition. In this light, the failure of the Burial scene in Bamberg to find its way into P49, paradoxical as it may appear—given Bamberg’s enormous contributions to the production of Junius pictures—is now subject to a principled account. Belonging to a different cycle that occurs later than the Adam and Eve series, the composition of P49 was simply denied access to the Touronian images for inspiration at all. Thus, the Bamberg image at issue would have been categorically unavailable for adaptation for P49. A question immediately arises: What would have determined this limitation of scope to the Adam and Eve cycle, thereby precluding whatever cycle occurred subsequent to it from being influenced in analogous fashion? It might be conjectured that the original Old Saxon Genesis episodes (see below) that provided the textual basis for illustration did not contain a narrative passage with a conceptual link—implicit or explicit—to the Bamberg picture. Put simply, the Old Saxon text involved would not have contained the story of Cain and Abel. In fact, the episode of Cain and Abel is an Old English original—rather than an adaptation from the Old Saxon—in the extant Old English Genesis in Junius 11, although a fragment of the Cain and Abel story, unillustrated, is transmitted elsewhere, namely, in the Old Saxon (Vatican) Genesis (verse lines 27–150; Doane 1991). Accordingly, the original (i.e., Touronian) artist had no need to visualize the Cain and Abel narrative for the Old Saxon Genesis episodes at his disposal. Thus, the Burial of Abel image found in Bamberg remained unutilized for the Touronian illustrated Old Saxon Genesis, a predecessor of the Old English/Old Saxon Genesis. Consequently, when composing pictures for subsequent episodes in the combined Old English Genesis (Genesis A2; see section 12.8 below), the Anglo-Saxon illuminator responsible for the project did not find a burial image available in the Touronian/Old Saxon model at hand to incorporate into P49 in the synthesized Old English work.

12.6 Auxiliary sources: The Utrecht Psalter and the Beatus Commentaries 

 281

12.6 Auxiliary sources: The Utrecht Psalter and the Beatus Commentaries 12.6.1 The Utrecht Psalter The Utrecht Psalter is often invoked as possible or probable sources for a small number of Junius pictures: Pii, P3.1, P3.2a, P3.2b, and P3.3. In addition, while conventionally not treated in relation to Utrecht, P7.1 and P11/P13 proved to be of prime importance in addressing the relationship between Junius and Utrecht (sections 4.2 and 6.2). This section reviews the roles that the Utrecht Psalter allegedly would have played or unexpectedly failed to fulfill, with a confirmation that Utrecht was merely of auxiliary significance in inspiring Junius. Despite occasional contentions to the contrary (Raw 1953: 68–69), there are no compelling reasons to believe that Utrecht provided a demonstrably solid basis for the founding picture of the pictorial system of Junius, namely, Pii (section 2.2). Among the three Utrecht images presented as possible sources for Pii, Utrecht 9r (Pl. 74), 59v (Pl. 79), and 76r (Pl. 80), the last item would look most similar to Pii. As criticized in depth (section 2.2), however, the similarities involved do not seem specific and individualistic enough to be accepted as cogent grounds for identifying Utrecht 76r (and also the other two) as a plausible source of inspiration. Indeed, the raising of the left leg (though slightly) in Utrecht 76r seems impressive, but this feature, as well as a cross nimbus and the personification of the winds, are such widely circulating motifs as to defy individual identification. Even if the clustering of these features may be acknowledged as significant, there are notable differences in evidence that would militate against privileging Utrecht 76r as a specific model of primary importance. The discrepancies include God seated on a globe, the winds not equipped with wings, the nonattendance of seraphim, the inward facing orientation of the flanking angels, and the absence of an architectural frame. In view of the insufficient specificity and individuality, and conversely, the wide currency of the features at issue, all one may be allowed to accept with empirical legitimacy is that Utrecht 76r may have been related to Pii as an instantiation of the Carolingian in general, and Touronian in particular, image of Christ in Majesty. The association in question is thus generic, concerning a class and its membership at the level of type, rather than a particular individual per se at the level of token. One may be reminded here that the Touronian exemplar of Christ in Majesty has been identified as a complementary source in the foregoing examination. Thus, whatever similarities one may recognize between Utrecht 76r and Pii are immediately subject to account within the explanatory framework proposed above, without devising an extra interpretive apparatus. In any event, lacking detailed specificities, the Utrecht image cannot be privileged as a particular model for Pii to the sacrifice of other comparable instances such as Vivian 329v. This conclusion has far-reaching implications going beyond one particular image, Pii. In view of the foundational status of Pii/P2 in Junius (section 12.2), the lack of con-

282 

 12 The Touronian Foundations of the Old English Genesis: A Synthesis

tributions by Utrecht to the core of the pictorial system may be adduced as a substantial refutation against the likelihood of Utrecht having served as a vital model for Junius at large, even if it should be acknowledged as an accidental stimulus in isolated cases. Unlike Pii, P3 seems to have been substantially inspired by Utrecht, specifically P3.1 by Utrecht 26r (Pl.  75), and P3.3 by a conjunction of Utrecht 37v/41v (Pls. 76 and 78), as discussed in section 3.2. As for the middle registers, P3.2a and P3.2b, there is no conclusive evidence for positing Utrecht as a model in precedence to, or at the least in complementation with, the Touronian Bibles. Thus, in contrast to the Touronian models, the effects of which are detected ubiquitously throughout P3, the Utrecht Psalter demonstrably bears on the first and last registers of the page, to the exclusion of the middle. Analogously, the contributions of Utrecht 26r to P3.1 are far from all-embracing; rather, they are restricted to the limited zone, the figure of Lucifer and the throne in the center, and the two processions of angels in the right half, whereas the left side of P3.1 displays no manifest indebtedness to the Utrecht image in question. Such limited effects on the pictorial zone sharply contrast with the across-the-board influence on the whole register that was exerted by Vivian 386v and San Paolo 310v. The two sources—Touronian and Utrecht—differ not only in the extensiveness, as just remarked above, but also in the depth of their effects. While the Utrecht image exhibits a momentous scene of a conglomeration of the ruler, the throne, and the double rows of angels to his right, a static configuration that might appear comparable to P3.1 in a way, no interaction between the ruler and each of the two rows of people is in play in Utrecht. By contrast, P3.1 is implicated in the multiple interactions between the pictorial components that are sequenced along the temporal axis: the upper procession of angels moving toward Lucifer in response to his beckoning, Lucifer himself in turn proceeding toward the palace, together with his followers obeying his invitation, and the lower row of angels approaching Lucifer to offer crowns inside the palace, a complex of dynamics that is foreign to the Utrecht image. Of striking interest, the Touronian models provide an exactly analogous horizontal movement in the corresponding register, with the central figure of Saul playing double roles corresponding to two distinct points in time. Inspired by the Touronian models, the Junius artist most likely would have practiced analogous pictorial dynamism and conflation of distinct acts into a single figure. As far as P3.3 is concerned, the falling crowd of people vividly depicted in Utrecht 37v/41v, especially the latter picture, might be adduced as convincing evidence for these Utrecht images having shaped the Junius representation of the falling angels. Indeed, the falling scene itself is an exclusive privilege of Utrecht, unknown to Vivian 386v2/ San Paolo 310v2, the major Touronian models for P3.1/P3.2a/P3.2b, thereby apparently casting doubt on the Touronian base for P3.3. Considered from a broader perspective, however, the formative role of the Touronian Bibles comes to light indubitably (section 3.2): the architectural structure of hell implicitly presupposed in P3.3 and explicitly represented in its variant images, P16.2/P17.2, and the particular arrangement of Satan and one of his former cohorts within the walled enclosure of hell represented in the last variant, P17.2, exhibit too specific correspondences to the depictions in Vivian 386v2 to

12.6 Auxiliary sources: The Utrecht Psalter and the Beatus Commentaries 

 283

be dismissed as pure contingencies. Accordingly, despite the superficial observation to the contrary, the Touronian underpinnings must be acknowledged no less firmly than the influence from the Utrecht Psalter. Of greater significance than the duality in sources of inspiration is a divergence in the nature of influence that the two models exerted. The Utrecht image concerns the overall gestalt of the falling crowd, whereas the Touronian ones bear on specificities of compositional details and conceptual parallels. As in P3.1, the influence of Utrecht images largely concerns relatively large assemblages of figures that tend to catch the eye even at a superficial level of casual observation. Rather than borrow such a static scene of conglomerated figures as it stands in the Utrecht original, the Junius artist reworked and incorporated this raw material into a dynamic succession of multiple interactive scenes in the strictly register-based framework in conformity with the Touronian style and conventions of representation. In view of the limited sphere of influence in terms of pictorial space and levels of composition, even where the Utrecht Psalter most likely would have inspired the composition of P3, its roles must be regarded as secondary to those played by the Touronian Bibles, which laid the foundations for visualization and therefore may be identified legitimately as the primary sources for the page (section 3.2). In identifying the ways in which the Utrecht Psalter inspired Junius, of more conclusive significance than the auxiliary role that it played in the creation of P3 may be a total lack of its appreciable contributions to the composition of P7.1 and P11/P13. Despite the common intention of, and the analogous compositional layout for, depicting in one piece a whole of earthly creations presided over by the Creator, neither P7.1 nor P11/P13 displays demonstrable indebtedness to P82v or other comparable Utrecht images. Utrecht 82v (Pl. 81) provides a panorama of heaven and earth, in which all God’s creatures are arranged symmetrically in multiple rows forming a single pyramid of the whole. By contrast, and extremely schematic and stylized, P7.1 represents a single frieze containing animals, birds, and plants on the ground, with no celestial bodies or water creatures above and below, despite ample space for representation. In short, P7.1 exhibits nothing remotely resembling the lively landscape full of living creatures that is presented in Utrecht 82v in a way that would strike one as naturalistic at first glance. In P11, one finds a crowd of animals inhabiting the sloping ground.146 The arrangement of these animals differs not only from P7.1, but also and no less markedly from Utrecht 82v. The animals in P11 are not arranged strictly in a frieze as in P7.1, but distributed naturally over the field. On close inspection, P11 diverges also from what might appear to be a naturalistic depiction—when contrasted with P7.1—in the Utrecht image. In fact, the whole array of animals, birds, and water creatures is subject to a structurally controlled arrangement in Utrecht 82v. Divided into three layers of frieze, the birds (top), beasts (middle), and water creatures (bottom) are symmetrically placed in

146 With a distribution pattern similar to multiple friezes, P13 looks somewhat closer to P7.1, as pointed out in section 6.2 above. Thus, the following discussion focuses on P11.

284 

 12 The Touronian Foundations of the Old English Genesis: A Synthesis

the right and left halves facing each other. Moreover, these creatures are represented in pairs for the most part. Such geometrically sophisticated designs characteristic of Utrecht 82v cannot be compared with the more naturalistic, irregular, plane–based or two-dimensional distribution found in P11. Thus, neither P7.1 nor P11 would have been demonstrably influenced by Utrecht 82v. On the contrary, the distinct manner of arranging animals in Utrecht 82v seems to have been avoided in both, although the three images at issue would have been predicated on the common motivation to represent a scene of God presiding over his creations. As substantiated at length in section 4.2 above, the appearance of only a single frieze of living creatures in P7.1 was due to the structural incompatibility of the format selected for P7.1—ultimately derived from the canon tables—with a multiplication of friezes executed as in Bamberg 7v2, Millstatt 9r, and Utrecht 82v. Therefore, the failure of Utrecht 82v to influence P7.1 was indeed motivated on formal grounds; the issue is not specific to the Utrecht model itself, but rather pertains to the abstract compositional scheme of multiple friezes as executed in Bamberg and Millstatt as well. The case of P11/ P13, however, concerns Utrecht 82v individually, as these Junius pictures are demonstrably modeled on Bamberg 7v2, an exemplification of multiple friezes, to the exclusion of the Utrecht candidate. Excluded here is not so much the pattern of multiple friezes per se as the specific instantiation of the pattern in Utrecht 82v. Thus, P7.1 resisted adaptation to Utrecht 82v (as well as to Bamberg 7v2) on generic grounds having nothing to do with individualities of that image. By contrast, P11/P13 evaded influence from the same Utrecht image presumably on account of its inherent properties, because the competitor, Bamberg 7v2, was successfully selected. As it turns out, then, P7.1 and P11/P13 are not implicated in the identical situation, although the same unsuccessful model, Utrecht 82v, is at issue in both cases. Accordingly, with a focus on P11/P13, one may be led to postulate two possible accounts—seemingly mutually exclusive—for the failure of Utrecht 82v to inspire P11/ P13 in defiance of all facilitating conditions. First, while having access to the Psalter, the artist remained unaware of 82v so that he missed an opportunity to consult it for P11/P13. This option would imply that the artist’s familiarity with Utrecht would have been fragmentary, superficial, and accidental, as he happened to catch a glimpse of the models for P3, but failed to do so for P11/P13 by chance. Alternatively, although the artist was familiar enough with Utrecht 82v, he dared not use it as a model for P11/P13. A question immediately arises: Why did he decide to reject the attractive image as a model, despite every good reason to take advantage of it? What are the intrinsic properties that blocked the artist from benefiting from Utrecht 82v? Lacking in the chosen model, Bamberg 7v2, but present in the rejected one, Utrecht 82v, is the sophisticated pattern of distributing paired animals symmetrically on the right and left halves of the panel, already pointed out. The pattern characteristic of Utrecht 82v is closely analogous to the face-to-face arrangement of Adam and Eve, below whom the array of animals would have been depicted. Thus, these apparently advantageous, unique properties would be extremely unlikely to have kept the artist from adapting the picture for P11/P13.

12.6 Auxiliary sources: The Utrecht Psalter and the Beatus Commentaries 

 285

In the absence of plausible accounts for this puzzle, the second account runs into an impasse. Eventually, one should be forced to settle on the first view adduced above: due to the artist’s limited familiarity with the Utrecht Psalter in general and his ignorance of Utrecht 82v in particular, he was unable to exploit this picture in composing the animal scenes in P11/P13. This reasoning, however, seems to be misguided, as it is predicated on a premise implicitly held that is open to question. According to the underlying assumption, Bamberg 7v2 and Utrecht 82v, both containing scenes of animals, would have been available as competitive resources for Junius pictures on an equal status. That is, the Junius artist would have been in a position to compare the two candidates in a fully objective way with a fair, open mind without bias in advance. Under such circumstances, the more elaborate design inherent in Utrecht 82v would have been more likely to be chosen. Such a strictly neutral basis of selection, however, would appear problematic and unrealistic. In light of the entire compositions of the first twenty-two pictures in Junius, the sources from the Cotton Genesis family in general and the Touronian branch in particular prevailed overwhelmingly. On such a markedly privileged basis of selection, it seems reasonable to assume that the Touronian sources were given first priority in consultation. Only thereafter were references made to other secondary reserve materials including the Utrecht Psalter insofar as deemed necessary to make a further search. Accordingly, inasmuch as the Touronian models were found to be exhaustively useful, no further action was taken to look at the Utrecht, even if the latter happened to be potentially more rewarding than the Touronian model. Thus, the Junius artist only occasionally would have turned to Utrecht as an ancillary repository of images, when he was obliged to do so by shortage of the traditional sources he was customarily working with, as with regard to P3. With respect to P7.1 and P11, by contrast, the artist had immediate access to the Touronian and Cottonian models that provided comprehensive and exhaustive materials on their own, and accordingly would have made extra use of the Utrecht Psalter redundant. The postulation of such a precedence relationship in the inventory of potential models may account for the nonselection of Utrecht 82v without necessarily implying the artist’s ignorance of the picture in question. From a broader perspective, the varying preference for and familiarity with possible sources, and the likelihood of their actual use, may be commonly reduced to their different positioning within the system of the working knowledge and craftsmanship that the artist put to practice in pictorial composition. Occupying the core of the artist’s cognitive system, the Touronian Bibles were immediately accessible, hence consulted first and then the Cotton Genesis—at the sacrifice of all others. Thoroughly knowledgeable about these central artworks at all levels of visualization, the illustrator was capable of exploiting them to the fullest extent. By contrast, the Utrecht Psalter was located on the periphery of the knowledge system in the mind, thereby resisting instant retrieval and accurate activation. In short, the artist lacked a competence of controlling the marginal material in a masterful way, on a par with the command he had at his disposal over the central apparatus.

286 

 12 The Touronian Foundations of the Old English Genesis: A Synthesis

12.6.2 The Beatus Commentaries As briefly mentioned in section 2.2 above, P2 displays significant compositional similarities to a number of the illuminated manuscripts of the Beatus Commentary. At issue is the scene of the “Woman clothed in the sun” (Rev. 12:1–18), in which the man-child is brought before God for rescue from the dragon. Of particular significance is the analogous placement of a nimbed young child at the center of the scene, standing to the left of the enthroned Deity. No less outstanding is the corresponding arrangement of this scene in the upper right corner of the entire visualization space (compare Pls. 2, 83, and 84). These resemblances seem too specific to be dismissed as purely accidental. Rather, it may be reasonable to posit that the two images at issue are genealogically related based on common ancestry. Originally having uncovered the common descent, Ohlgren (1975: 41–44) formulated a contact-based account by asserting that Junius was influenced by the Beatus Commentaries. With full commitment to the conventional view that P2, like all other pictures in Junius 11, was produced in Late Anglo-Saxon England, Ohlgren (1975: 44–46) was forced to conjecture, in the absence of conclusive evidence, a close contact between England and Spain in the tenth to eleventh centuries, an assumption that was criticized by Broderick (1978: 78). By marked contrast, the claim that the first twenty-two pictures—particularly P2 and P3—were composed at Tours in the mid-ninth century may be capable of offering a credible account of the similarity between P2 and the man-child in the Beatus Commentaries. Since Ohlgren’s (1975) findings, the knowledge and understanding of the Beatus Commentaries have been dramatically widened and deepened thanks to Williams’s (1994–2003) copious corpus of illustrations. In this light, it may be appropriate here to dwell on the contact-based relation between the composition of P2 and the Beatus Commentaries, and explore some empirical implications that this connection may suggest for further investigation. Among the twenty-six extant illustrated Beatus Commentaries catalogued in Williams (1994–2003.1: 10–11), thirteen contain relevant images, as listed below on the basis of a classification into two subsets. That is, the scene of the man-child before God may be divided into two variants, based on whether or not an angel is depicted to the right of God as well, in addition to the one invariably standing behind the child at the left. Two groups of the man-child before God a. God flanked by angels (closer to P2): five manuscripts – Morgan (M; New York, Pierpont Morgan Library, MS M.644), fol. 153r (Williams/ Shailor 1991; Williams 1994–2003.1: pl. 5; Williams 1994–2003.2: pl. 62); dated 940–945. – Valladolid (V; Valladolid, Biblioteca de la Universidad, MS 433), fol. 131r (Williams 1994–2003.2: pl.  193; Beato de la Universidad de Valladolid 2000–2002); dated 970.

12.6 Auxiliary sources: The Utrecht Psalter and the Beatus Commentaries 

 287

– Urgell (U; Urgell, Museu Diocesá de La Seu d’Urgell, Num Inv. 501), fol. 141r (Williams 1994–2003.3: pl. 53; Codex Urgellensis 1997–2002); dated 975–1000. – Facundus (J; Madrid, Biblioteca Nacional, MS Vitrina 14-2), fol. 187r (Pl.  83; Códice de Fernando I 1994; Williams 1994–2003.3: pl. 296); dated 1047. – Silos (D; London, British Library, Add. MS.  11695), fol. 148r (Williams 1994– 2003.4, pl. 284b; Códice de Silos 2001–2003); dated 1109. b. God at the extreme right, without an angel to the right: eight manuscripts – Girona (G; Girona, Museu de la Catedral de Girona, Num Inv. 7(11)), fol. 172r (Williams 1994–2003.2: pl. 342); dated 975. – San Millán (A2; Madrid, Real Academia de la Historia, Cod. 33), fol. 160r (Williams 1994–2003.3: pl. 128; Beato de San Millán 2002); dated 975–1000. – Escorial (E; Escorial, Biblioteca del Monasterio, &. II.5), fol. 105r (Williams 1994–2003.3: pl. 195); dated c. 1000. – Saint-Sever (S; Paris, Bibliothèque Nationale, Ms. lat. 8878), fol. 159r (Pl.  84; Códice de Saint-Sever 1984; Williams 1994–2003.3: pl.  432); dated mid-11th century. – Osma (O; Burgo de Osma, Archivo de la Catedral, Cod. 1), fol. 117v (Williams 1994–2003.4: pl. 43); dated 1086. – Turin (Tu; Turin, Biblioteca Nazionale Universitaria, Sgn. I.II.1), fol. 132r (Williams 1994–2003.4: pl. 166b); dated 1101–1125. – The Rylands (R; Manchester, John Rylands University Library, Latin MS 8), fol. 143r (Williams 1994–2003.5: pl. 77; Códice de Manchester 2001); dated 1175. – Las Huelgas (H; New York, Pierpont Morgan Library, MS M.429), fol. 102r (Williams 1994–2003.5: pl. 393; Beato del Monasterio 2004); dated 1220. Needless to say, the images with a double representation of angels—the first group— correspond to P2 more closely than those with a single angel. Of special interest here is the distribution pattern of these two variants. The exact formulation of the distribution eventually depends on the specific relationship that one postulates between Subbranches IIa and IIb—internal to Branch II—in textual and pictorial terms. Specifically, the two subgroups within Branch II—IIa and IIb—may be regarded as collateral, derived from a common archetype (Neuss 1931.1: 111; as illustrated in Williams 1994–2003.1: 22; Klein 1976; as illustrated in Williams 1994–2003.1: 23). Alternatively, Subbranch IIb may be considered subordinate to Subbranch IIa, having branched off from the latter as a secondary subgroup (Williams 1994–2003.1: 26). In light of the genealogy of the illustrated Beatus Commentaries posited by Neuss and Klein (see above), the representation of the enthroned God flanked by an angel on either side may be characterized as limited to Subbranch IIa. Alternatively reformulated, following Williams’s family tree (Williams 1994–2003.1: 26), the symmetrical representation of angels is internal to Branch II, but does not extend down to Subbranch IIb. By either formulation, however, it remains the same that the double presence of angels is confined to the set of Branch II excluding Subbranch IIb; in other words, the double

288 

 12 The Touronian Foundations of the Old English Genesis: A Synthesis

angels are absent in Subbranch IIb as well as Branch I in its entirety. In short, the doubling of angels constitutes an extremely restricted feature in genealogical terms, in opposition to the single presence that is a default manifestation elsewhere. Hence, it may follow that the appearance of an angel on the right of God must have been an innovation added to the conventional depiction of a single angel on his left. Given that the genealogical grouping is motivated independent of the issue of the single versus double angels, the correlation of this feature to the genealogy must be regarded as significant and worthy of explanation from a genealogical perspective: it is far from an accidentally variable, idiosyncratic trait to be simply dismissed out of account. Of particular interest about the genealogy-sensitive distribution of the double angels is its subsumption under the demonstrable sphere of influence by the School of Tours in general on the Beatus Commentaries. The Touronian influence is limited to Branch II, common to its subgroups IIa and IIb alike. Thus, the Touronian effects in general are visible in a wider range of Beatus than the double angels. This is only natural, given that the influence as a whole from the Tours School is an aggregate of manifold features that vary individually in the scope of their influx. Actually, there is ample, substantial evidence that the Touronian Bibles provided inspiration to the illumination of the Beatus Commentaries and beyond, as testified by, among others, the banded backgrounds (compare with Grandval 449r and Vivian 415v, frontispiece to the Revelation; Williams 1994–2003.1: 62), the decorated initials (e.g., Morgan Beatus 10r, compared to a Bible with prologs and capitula [Bibel mit Prologen und Capitula], from Tours; Zurich, Zentralbibliothek, Ms Car C1, fol. 359v; Williams 1987: 198–199; 1994–2003.1: 78), and the composition of the Majesty and Cross pages in the Girona Beatus 1v and 2r, corresponding to Christ in Majesty in Vivian 329v and the Lamb in Majesty in Bamberg 339v, respectively (Williams 1987: 200–201; 1994–2003.1: 62–64). Since the earliest extant Beatus manuscript of Branch II is the Morgan Beatus, dated 940 to 945 (Williams 1994–2003.1: 10, 62), the terminus ante quem for the inception of the Touronian influence may be put at the 940s, as convincingly reasoned by Williams (1994–2003.1: 76). In light of the manifold indications of the Touronian influence on the Branch  II Beatus Commentaries, one may pursue two lines of interpretation with regard to the similarity between P2 and the man-child scene, committed to the hypothesis that the resemblance stems from the common source involved. First, as claimed by Ohlgren (1975), the Beatus illustration would have served as a model for P2. While the inspiration going from Tours to northwestern Spain is fully substantiated, there seems to be no independent evidence for the opposite direction. In general terms, however, an absolutely unidirectional path of influence would be scarcely as plausible as a bidirectional one, even if normally far from symmetrical in strength relative to directionality. In any event, the demonstrable influence of the Touronian Bibles on the Beatus Branch II family would render the contrary course of influence more feasible than direct contact between Spain and Late Anglo-Saxon England, specifically in terms of manuscript illumination that Ohlgren has to postulate on inconclusive grounds. Thus, according to this view, P2 would have been inspired by the Beatus scene of the man-child (part of

12.6 Auxiliary sources: The Utrecht Psalter and the Beatus Commentaries 

 289

the Woman Clothed in the Sun) characteristic of Subbranch IIa, as exemplified in the Morgan Beatus 153r. Alternatively, in the absence of substantive Beatus–to–Tours influence, one may be led to conjecture that the resemblance at issue took the opposite direction of transference: it may have been brought from Tours to Asturias, along with other Touronian features of manuscript illumination. Since, of the two variants mentioned above, the one with two angels is compositionally closer to P2, this must have been the new image transferred from Tours to Asturias and established as a distinctive innovative feature of Subbranch IIa by displacing the traditional one with a single angel. These competing accounts require evaluation by exploring their empirical implications. At issue are two differing entailments that they each may make, derivational accountability and chronological plausibility. Derivational accountability concerns how convincingly the compositional similarity can be accounted for in terms of image transferal. In the first interpretation, the entire scene in P2 of the nimbed child standing before the enthroned Deity flanked by angels may be viewed as having been borrowed from Beatus. Thus, the substance of this view is close to Ohlgren’s claim, although the recipient and channel of contact postulated are markedly different. By contrast, according to the second account, inasmuch as only the presence of two angels in P2—as at first glance—groups Subbranch IIa with P2 in opposition to other Beatus classes, it may follow that the nimbed child standing before the enthroned God, which is shared by the entirety of the Beatus images as well as P2, would not have been part of transferal. Or to reformulate, the transferal would have affected the shared part vacuously. It follows as a matter of course that this trait must have been as inherent in the Beatus illustrations in general as in P2, immune to borrowing in explicit form. Thus, the commonality between P2 and Beatus at large of the motif of the man-child in front of the enthroned God would have to be left unaddressed in the second view. In order to account for the common representation of the man-child standing before God, still another source must be invoked. Minimizing ad hoc nature incurred, one might appeal to a still earlier manuscript of Beatus (of Branch I) as a model. In view of the long tradition of contact between Spain and the monastery of St. Martin at Tours (from the eighth century on; Williams 1987: 201–202), such a derivation might be conceivable, although lacking enough empirical specificity to claim immediate acceptability. A further possibility, apparently no less ad hoc, would be to invoke a common source behind P2 and Beatus (archetype) without positing direct contact between them at this earlier stage of production, namely, the middle ninth century. This line of thinking would seem methodologically far from objectionable, in the absence of evidence for specific artistic contact (cf. Broderick 1978: 78). In summary, the movement from northwestern Spain to Tours would appear to be slightly more realistic than the contrary direction of contact, which requires either a still earlier contact or a vague common source that would have been available in both Asturias and Gaul. The above assessment, however, is premature; it needs a radical reappraisal by broadening the perspective. The tentative evaluation at issue rests on the premise, as

290 

 12 The Touronian Foundations of the Old English Genesis: A Synthesis

stated explicitly in the preceding paragraph, that the distinction between single and double angels in the Beatus images is an isolated property that distinguishes between Subbranch IIa and others. On closer inspection, it turns out that the difference between single and double angels significantly correlates to other pictorial distinctions. In the single-angel type, God is usually shown receiving or about to receive the child, as is typically represented in Escorial, Osma, Turin, and Las Huelgas: his hands (or hand) are extended specifically to the child who, in turn, is approaching God with both hands raised as for help; correspondingly, the distance between God and the child is minimal. In contrast, the double-angel type represents God addressing the child and/or the angel behind it: God points his hand toward the angel standing behind the child, and separated from each other at a distance, neither God nor the child displays marked gestures suggestive of a physical contact about to take place between them. As a depiction of God addressing, the second variant is closer to P2 than the first. In addition to the pictorial content, the framing and spatial arrangement constitute correlative differences as well. In contrast to the double-angel type, the single-angel one is not always located in the upper right corner of the page, as in Escorial. The double-angel type always delineates the man-child scene by framing the upper right corner of the page allocated to it, whereas the single one does not invariably do so (e.g., San Millán and Escorial). Thus, the contrast of the single versus double angels proves to be part of the larger clustering of multiple pictorial distinctions. As observed above, the scene of God addressing, which demonstrably associates with the presence of two angels, corresponds more closely to P2 than that of God receiving the child, which is inseparably linked with a single angel. While God’s act of addressing in P2 perfectly fits the narrative content in the Fall of the Rebel Angels (section 2.1), the corresponding Beatus image seems less suited to the Revelation context than the variant with a single angel, in which God is shown being about to take on the endangered child himself for rescue. Judging by varying narrative fitness and distribution, the inference may be warranted that the scene of God addressing and flanked on both sides by angels would have been native to P2, rather than the Subbranch IIa family of the Beatus Commentaries, which must have acquired it secondarily from outside. It may be worth reminding with emphasis here that the composition of P2—including the two seraphim immediately beside God—was most likely predicated on Vivian 423r, as substantiated on a solid empirical basis (section 2.2), and, no less importantly, that the representation of Lucifer as a small child, the pivotal figure of the scene, would have been fully motivated on internal conceptual grounds (section 2.1). Thus, the derivational precedence of P2 over Beatus—rather than the converse—with regard to the representation of the two angels flanking the addressing ruler and other associated pictorial features, as substantiated above, is fully compatible with the independently established descent of the pre-Junius Old Saxon Genesis episodes from the Touronian Bibles in general and that of P2 from Vivian 423r in particular. A widened perspective, therefore, may choose the direction of influence moving from Tours to Asturias: the man-child scene in which God addresses the child and/or

12.6 Auxiliary sources: The Utrecht Psalter and the Beatus Commentaries 

 291

the angel behind it, and in which another angel appears to the right of God would have been modeled on the image manifested in P2, through the Carolingian, and notably Touronian, contact that introduced multifold artistic influence to Asturias. Moving on to the issue of chronological plausibility, the first view—influence from Beatus on Junius—proves to be infeasible. Since the illustrated Old Saxon Genesis episodes must have been produced in the mid-ninth century in or near Tours as assumed and substantiated throughout the foregoing examination, the innovative Beatus scene necessarily would have been available by that date to make its way into the Touronian/Old Saxon Genesis. This in turn would entail that the terminus ante quem for the emergence of the innovative Subbranch IIa feature—the depiction of two angels before and behind God and other associated features uncovered above—must be the midninth century. This date would appear too early to be accepted in light of Williams’s carefully reasoned dating of Branch II mentioned earlier. Since the earliest exemplar of Subbranch IIa is the Morgan Beatus, dated to the 940s, it would take minimal extra assumption to reason that Subbranch IIa came into being not long before that date.147 That is, it would be far-fetched to assume that Branch II manuscripts would have been produced already in the mid-ninth century, almost a century before the first attestation ever known. By contrast, the dating of Branch II at the 940s poses no problem for the second view in terms of historical plausibility. The archetype of the Branch II variant of the manchild scene originated from P2 in the Old Saxon Genesis composed at Tours in the midninth century. Subsequently, at whichever date before the 940s—most likely not long before the terminus ante quem—the pre-Junius image or its variant would have been transferred to Asturia along with other Touronian artistic features probably through a copy (or copies) of a Touronian Bible (and/or related works; cf. Williams 1987: 201–202). Thus, the assumed chronological trajectory requires no special pleading going excessively beyond empirical substance. Accordingly, in total agreement with the decision based on derivational plausibility, the chronological assessment decides on the second view that the similarity between P2 and the Subbranch IIa Beatus image of the manchild before God would have arisen through inspiration coming from Tours to Asturias in the second quarter of the tenth century. Now that the Touronian inspiration for the Subbranch IIa Beatus Commentaries has been established beyond reasonable doubt, one should address its empirical implications to situate substantially the Old Saxon illustrations of the Old English/Old Saxon Genesis in the Touronian/Cottonian tradition. That is, one would have to accept that the Touronian/pre-Junius innovation as embodied in P2 would have been incorporated into the common stock of the Tours School—rather than having remained as a parochial trait 147 Compare Williams’s remarks as follows: “Whenever our Touronian Bible, or Bibles, arrived, the fact that Leonese manuscripts [= Branch II] were affected first must mean that it was by way of Astúr-Léon rather than Castile or Aragon that this Carolingian school contributed to the revolution in Hispanic illumination, and then not before the second quarter of the tenth century” (Williams 1987: 202).

292 

 12 The Touronian Foundations of the Old English Genesis: A Synthesis

of one of its outlying manuscripts, an illustrated Touronian copy of Old Saxon Genesis episodes—so as to further enter into Beatus as part of Touronian book illumination. This would imply, much as in the case of P41/P44 (section 12.4), that the Touronian/Old Saxon/pre-Junius Genesis contributed to enriching the Touronian tradition by supplying an innovative image for common use.

12.7 Subsequent alterations This section summarizes major alterations that would have been added to the Touronian/Old Saxon original in the course of transmission, culminating in the extant Junius manuscript. Operating on the three major figures, God, Adam, and Eve, respectively, these alterations rendered the Touronian identity of the Junius pictures opaque. Demonstrating the derivative nature of these subsequent changes will restore the Touronian underpinnings of the Junius images more substantially.

12.7.1 The scroll/scepter and the cross nimbus Resembling a scroll or scepter but defying either categorization is the rather enigmatic object that God is shown holding in his hand in the extant Junius manuscript, which has been referred to as a scroll/scepter throughout the foregoing discussion (sections 2.1 and 2.2). As detailed in the previous sections, this item, which made its first appearance in Pii as an attribute of God enthroned, would have arisen through a reconfiguration of the original combination of a book placed on the left lap and a scepter stacked on top of the book, held in the hand at the junction of the two objects. In the meantime, it was proposed that this image of God enthroned in the frontispiece was derived from two models on a complementary basis, Vivian 423r being primary and the Touronian exemplary representation of Christ in Majesty being secondary. Only by invoking these two sources in conjunction, it was claimed, may an array of significant features of the Creator be subject to principled accounts. Of paramount importance in appealing to Christ in Majesty is the marked positioning of Christ’s left leg higher than the right, distinctively characteristic of the Touronian representation of the scene at issue. Since the ruler is shown seated on the throne with his feet level in Vivian 423r, their conspicuously uneven arrangement in Pii may be understood by deriving it from the Touronian exemplar of Christ in Majesty, which embodies exactly this trait. The allegedly original configuration of a book and a scepter, however, is not in evidence in Pii. Instead of the book, one finds a board-like object placed on the left lap below the scroll-scepter. It may be conjectured that this rod-like complex standing on the left lap was a reflex of the original superposition of a scepter on top of a book. Specifically, the bottom section—a board-like, inarticulate part—would have emerged as a consequence of a decategorization as a book, a cognitive obscuration resulting in a denial of

12.7 Subsequent alterations 

 293

precise categorization. In exchange for this conceptual loss, the upper part—originally a long scepter, as manifested in Vivian 423r—would have absorbed part of the identity of the obscured element—a book—to be reconfigured as something conceptualized as a scroll as well. Meanwhile, the originally long scepter had been subject to shortening, not least because holding such a long rod on top of something else would have struck as unnatural and clumsy in terms of physical posture and visualization. Concomitantly, this shortening would have been found in harmony with the emergent recategorization of the upper part as a scroll and the advancing decategorization of the lower part as a book. Yet the original categorization as a scepter would have hung on, rather than giving way completely to the reanalysis as a scroll. Hence, the ambivalent status of a scroll/ scepter would have been reached in compromise, as witnessed in Pii and subsequent pictures in Junius. Such complex processes of decategorization/recategorization would have been prompted by the pose of the left leg raised higher than the right, characteristic of the Touronian image of Christ in Majesty (section 2.1). The raising of the left leg in conformity with the Touronian exemplar would have lifted the top of the scepter too high, thereby demanding a radical shortening of the whole complex for the sake of physical stability and compositional balance. Having overall effect on the composite item in its entirety, this demand for shortening affected the lower part as well by obscuring the existentiality of the book itself. In brief, it may be claimed that the repositioning of the left leg higher than the right would have made it difficult to retain the original juxtaposition of a book and a scepter along the vertical axis, thereby promoting the compaction and unification of the two items into a scroll/scepter, and consequently obscuring the dual models behind the image of God enthroned in Pii (see further below). The evolutionary trajectory for the emergence of the scroll/scepter proposed above receives empirical corroboration from the representation of Christ in Majesty in Boethius (Paris, Bibliothèque Nationale, Ms. lat. 6401), 158v (Pl. 73). In it, Christ holds “a book and symbol of authority [= scroll/scepter according to the present designation: S.S.] in His Hand” (Temple 1976: 59). Thus, an analogous composite structure is involved. However, the identity of the lower part—the book—is beyond question here, and the symbol of authority—the upper part—is much shorter (by half) and thicker (more than double) than the scroll/scepter in Pii. Specifically, the unbanded portion of the scroll/ scepter in Junius is lacking in Boethius, whereas the book underneath is fully depicted as such. In overall synchronic terms, then, while the original book has lost its identity in Junius, the scroll/scepter (banded in the upper, plain in the lower) corresponds to the composite of the book (below; unbanded) and the symbol of authority (above; banded) in Boethius, which retains the original dual composition intact. In light of the structural isomorphism involved, it may be justified to postulate the common configuration of a scepter on top of a book underlying both representations. In contrast to the neat correspondence of the items held in the hand, the posture of the left leg is divergent between the two figures of God: the figure of Christ in Boethius does not have his left leg raised, naturally enough, given that this is not a work of the

294 

 12 The Touronian Foundations of the Old English Genesis: A Synthesis

School of Tours. At this point, the assumption presented in the preceding paragraph proves to be explanatory. The raising of the left leg would have made it difficult to keep the original composite structure intact because of the physical inconvenience and compositional clumsiness of holding a resultant excessively tall object. The greater demand for shortening the complex item eventually would have led to a decategorization of the book, and the concomitant emergence of the scroll/scepter in Junius. In the absence of the initial cause for lifting the object too high, the excessive requirement for making it compact did not materialize for Boethius. Therefore, the original duality in composition persisted there. According to the proposed view, the conjunction of a book and a scepter underlying the scroll/scepter stems from the dual sources of Pii—Vivian 423r that supplied a scepter and the Touronian image of Christ in Majesty that provided a book. The independent existence of the fully articulated dual composite in Boethius confirms empirical credibility of the hypothesis. The retention of structural duality in Boethius on the one hand and the obliteration of it in Junius on the other may be explained by whether Christ is shown with his leg raised, namely, with which source the image of Christ in Majesty is affiliated, Tours or somewhere else. Since the two models genealogically converged in the composition of Pii, it may follow that Boethius would have obtained the book/ scepter composite secondarily, rather than forging it on its own. While details await future investigation, it may be tentatively but plausibly assumed that the Anglo-Saxon artist responsible for the Boethius illustration in question (Temple 1976: 59) picked up the configuration in the last quarter of the tenth century, when it must have been in currency in its own right—no longer a privilege of the School of Tours—for more than a century after the inception at Tours. This may be regarded as another instance of Junius innovation that was transferred from Tours to the outside (see also sections 12.4 and 12.6.2). Given that the unmarked posture is to put both feet level, the strict conformity to the typical figuration of Christ in Majesty characteristic of the School—the raised left leg—may suggest that the artist was deeply committed to the Touronian exemplar. Since Vivian 423r, a presentation miniature as it is, naturally does not instantiate the pose of Christ in Majesty, particularly in regard to the raising of the left leg, the manifestation of this feature in Pii/P2 may testify the artist’s profound indebtedness to the Touronian practice. It would be a short step to infer that the artist was a member of the School of Tours. The dual sources of God enthroned in Pii do not exclusively concern the origin and development of the scroll/scepter, as treated above. Another issue of interest has to do with a cross nimbus God carries in Junius. As Kessler (1977: 26–27) points out, “In the Cotton [= San Marco: S.S.] and Klagenfurt [= Millstatt: S.S.] manuscripts the Creator is identified as Christ by his cross nimbus. This Christological feature is absent from all the ninth-century manuscripts and therefore seems not to have been in the model.” Accordingly, the categorical absence of a cross nimbus in the Touronian Bibles and its presence in the purportedly Touronian-based Junius would appear contradictory at first

12.7 Subsequent alterations 

 295

glance. Either Kessler’s generalization on the persistent avoidance of a cross nimbus as an attribute of the Creator would require qualification, or the proposed derivation of Junius pictures from the Touronian models would need additional refinement by postulating a later introduction of a cross nimbus, a kind of Anglo-Saxonization or an undoing of Touronian associations. There is in fact a third solution available, the one that would count as minimal in cost of necessitating further assumptions, and that accordingly should be adopted as optimal and simplest among the three competing options. The appearance of a cross nimbus in the Creator in Junius would be explained as an inevitable consequence of God enthroned (Pii) being modeled on the image of Christ in Majesty, which is provided with a cross nimbus even in the Touronian Bibles such as Vivian 329v (Pl. 30). Therefore, the use of a cross nimbus in subsequent pictures in Junius does not mean that it necessarily resulted from Anglo-Saxonization after having been transferred to England. The cross nimbus likely would have been used from the very beginning at Tours when a Touronian/Old Saxon predecessor of Pii was originally created. Seen from a different perspective, since the Touronian image of Christ in Majesty was postulated above as a source for Pii conceptually independent of the appearance of a cross nimbus in it and on subsequent pages in Junius, such depictions should be adduced as support for, rather than objection against, the proposed hypothesis that Pii was based on the Touronian exemplar of Christ in Majesty.

12.7.2 The recomposition of the Temptation of Adam as the Temptation of Eve If all Temptation scenes in the extant manuscript of Junius demonstrably concerned Eve to the exclusion of Adam, this would detract from plausibility and legitimacy of situating the artwork in an environment in which the Old Saxon original of Genesis B (or at least part thereof) was readily accessible and familiar enough to demand accompanying illustrations for a better understanding and greater pleasure, as hypothesized in this book. As fully substantiated in section 7.1, however, there are empirical grounds to assume that P20 and P24 were intended as representations of the Temptation of Adam according to original conceptualization. Therefore, the continental foundations of the Junius pictures up to P46, including P20 and P24 in their original form, are hardly open to doubt and outright rejection. Confronted with the indubitable identification of Eve being tempted in the extant depictions in P20 and P24, however, one must postulate a subsequent recomposition at some stage of transmission from the initial production to the final fixation in the current manuscript. Thus, the transfiguration from the Temptation of Adam to that of Eve must have occurred in the course of copying and editing in the pre-Junius period on Anglo-Saxon soil (see below). The motivation for such pictorial transformation can be readily understood. Ultimately at stake would have been the persistent force of the biblical tradition demanding conformity and assimilation, and conversely a weak counter power of the vernacular

296 

 12 The Touronian Foundations of the Old English Genesis: A Synthesis

Genesis in the face of overwhelming biblical authenticity. It should be noted in this connection that the revisualization of the Temptation of Adam as that of Eve in accordance with the biblical tradition would not immediately contradict the narrative content of Genesis B. The reinterpretation along the line of the Book of Genesis would simply skip and remove existing direct references to the Temptation of Adam in the text, and address the Temptation of Eve exclusively. Accordingly, the resultant pictorial narrative reconstituted in this fashion would hardly conflict the Old English/Old Saxon Genesis overtly. In view of the minimal likelihood of the ensuing conflict with the text, one may conjecture, even in the absence of a strong intention to suppress what might have appeared to be an incorrect episode of the Temptation of Adam, this prospective lack of manifest confrontation with the text would have convinced a moderate-minded copyist to change Adam to Eve without much hesitation. In other words, the figural change of Adam to Eve in the Temptation scene may have been executed only as a modest and innocent manipulation. Nonetheless, the oddity would have remained inerasable that the important passage of the devil attempting to deceive Adam escaped visualization. As a result of the reinterpretation, there would have arisen three consecutive pictures dealing with the Temptation of Eve: P20, P24, and P28. It should be borne in mind further that there must have been a few more pictures of the Temptation of Adam on the lost pages (section 7.1). This would have resulted in at least five pictures of the Temptation of Eve (P20, P24, P28 + at least two others lost), sequenced as follows: A hypothetical series of the Temptation scenes P20, PX, PY, P24 (originally Temptation of Adam)

P28 Temptation of Eve

P31 Fall of Adam and Eve

PX, PY would have appeared somewhere between the current pages 22 and 23. (See Table 2, section 1.1, for reference.) As observed in section 7.1 above, the gender transformation would have consisted in straightforward additions of a few physical attributes characteristic of a female, the nipples and the long hair. These strictly local operations did not incur further adjustments in other parts of the pictures at all, whether within the same figure of Adam/ Eve—the orientation of the body and the hand gesture—or beyond. Thus, the alterations added later must be characterized as superfluous, failing to reach the inner dimensions of interpictorial relations that are wholly and distributively loaded with the signification of Adam—rather than Eve—being tempted. That the copyist thus limited the change to a few obvious physical traits in the upper body of the original figure of Adam without duly exploring structural consequences may suggest that he scarcely would have been remotely familiar with the network of intervisual relations underlying individual features accessible to casual observation.

12.7 Subsequent alterations 

 297

In conclusion, the vulnerability to the prevalent biblical tradition to the sacrifice of specifics unique to the vernacular Genesis, and the partial, superficial alteration executed would lead one to infer that this revisionist–copyist, however much he was intent on revision, was situated far removed from the native environment in which the Old Saxon Genesis was open to reception in the original and the illustrations to it were provided for a correspondingly enriched appreciation.

12.7.3 The spindle whorl in P45.2 Since all members of the Touronian Bibles including those nonmanuscript descendants derived from the Touronian pedigree, such as Hildesheim, depict Eve nursing a child—rather than engaged in spinning—in the Labor scene (Table 12, section 12.2), the appearance of what would appear to be a spindle whorl in Eve’s hand at the Expulsion (P45.2) in Junius 11 would seem hardly compatible with the purported Touronian identity of Junius, and possibly might go so far as to put the claim under suspicion. In contrast, the association of Eve with spinning at Labor is visualized in a minority of the Cotton family, namely, San Marco and the Hortus Deliciarum, in addition to Junius. Such a distribution pattern might be interpreted as suggesting a Non-Touronian/Cottonian character of Junius in this regard. Unlike Junius, however, San Marco (Pl. 51) and the Hortus Deliciarum (Pl. 70; Green et al. 1979: pl. 13, no. 24) have a whole set of spinning tools depicted—a distaff, a spindle, and a whorl. Moreover, Junius is inconsistent in its treatment of the whorl: while it appears in P45.2, it is absent in P46, the picture that would have been derived in part from the Labor scene (section 11.2). Thus, only the least visible component of the spinning implements—a spindle whorl—was chosen for representation in Junius, as if to mask the resultant connection between Eve and spinning at Labor that would displace the Touronian association of Eve nursing a child. Furthermore, the depiction was executed with apparent inconsistency in that it is displayed in only one of the two consecutive scenes, as mentioned above. In light of the minimal amount of prominence attached to the tool and its existential instability, it may be warranted to infer that the whorl was a later and nonsystematic addition—rather whimsical at that—to the original illustration in which spinning tools would have been totally absent on the continent at the beginning. Therefore, the inconspicuous presence of the spindle whorl in P45.2 proves to be fully compatible with the identification of Junius as Touronian in origin.

298 

 12 The Touronian Foundations of the Old English Genesis: A Synthesis

12.8 On contextualizing the production of an illustrated manuscript of the Old English Genesis: implications and conjectures While this formal and cognitive study is primarily concerned with issues internal to the pictorial organization of the Old English Genesis in its synchronic and diachronic dimensions, it has richer implications that go beyond the problematics initially defined and fully addressed. Accordingly, it may be appropriate at this juncture to examine some of the empirical consequences, and, on that basis, to make reasoned inferences regarding the context of the production of an illustrated manuscript of the Old English Genesis (including its Old Saxon prefiguration) prior to its fixation in Junius 11. That is, drawing and elaborating on Doane’s reconstruction (1991: 43–54; 2011: 65–70, 81; 2013: 31–41), one may well outline, necessarily on a conjectural note, a prehistory of the Old English Genesis as it was shaped by being merged with its Old Saxon counterpart, before its eventual materialization in the surviving manuscript. As the foregoing investigations have uncovered, one is struck by the artist’s profound knowledge of the underlying system and principles of pictorial organization and composition, including their varying exemplifications in actual practice and their rich potential for experimentation (sections 12.2 and 12.3). Situated firmly in the artistic tradition that the Tours School artisans had cultivated, he was fully qualified for the innovative and imaginative execution that he made of such inherited artisanship. Of related significance is his ranked indebtedness to a variety of external sources, whereby the Touronian models (as embodied by the Touronian Bibles) outrank the Non-Touronian/Cottonian ones (as reflected later in the San Marco mosaics; section 12.3), which in turn prevail over the Non-Touronian/Non-Cottonian ones (as represented by the Utrecht Psalter; section 12.6.1). Such finely graded familiarities with the specific external sources in the artist’s cognitive system prove to be superior in explanatory power than the unordered assemblage of diverse Carolingian influences and beyond that were conventionally invoked for comparison on the surface (cf. Raw 1976: 146; Doane 2011: 66). All these combined make it most plausible to infer that the original artist responsible for composing the twenty-two pictures—or at least a major subset of them—was an active member of the School of Tours. The substantial influence of the Vivian and San Paolo Bibles further indicates that he would have been trained at an advanced stage of the artistic tradition; he would have acquired, through a dedicated apprenticeship, a masterful command of the Touronian artisanship richly accumulated, which, however, suddenly came to an end as a collective knowledge soon after 850 due to Viking invasion. Thus fully equipped and working at (or near) Tours in the mid ninth century, he would have composed these illustrations as accompaniments to the relevant episodes of the Old Saxon Genesis not long after its emergence. Given that the composition of the Old Saxon Genesis can be dated to near the end of the second quarter of the ninth century on literary and historical grounds (cf. Doane 1991: 46; 2011: 81), the dating of the illuminated

12.8 On contextualizing the production of an illustrated manuscript of the Old English Genesis 

 299

Old Saxon manuscript to the mid ninth century, as postulated primarily on its pictorial basis, seems chronologically consistent and most feasible in historical reality. The alternative view of the Old English Genesis illustrations having been produced for the first time on Anglo-Saxon soil—at whichever date—confronts the insurmountable difficulty of accounting for the deep working knowledge of the Touronian tradition that the illuminator demonstrably had at his disposal, whether with respect to the external models or system-internal motivations. The conceptual implausibility is compounded further by a postulation of Late Anglo-Saxon illumination (cf. Ohlgren 1975: 63), as the greater distance in temporal terms thereby entailed will simply add to the initial infeasibility. Such a full revival of the past artisanship long fallen out of practice seems extremely unrealistic, particularly in the absence of comparable contemporary works elsewhere. The construction on the continent of an illuminated manuscript of Genesis in the vernacular would have been inspired in part by the Cotton Genesis tradition itself, as with the Millstatt Genesis brought into being a few centuries later (section 1.2). Also influential was the inherent alignment of illustrations with accompanying verse texts (tituli) in the Touronian Bibles. Thus, the versification of images, and conversely the visualization of verses, inhere in the primary models of the Old Saxon Genesis. Moreover, standing on its own, the Old Saxon Genesis epic (or, more likely, a collection of its individual constituent episodes; cf. Doane 1991: 42) was no longer subject to the strict limitation of illustration to a single page per book, as bindingly characteristic of the one-volume Touronian Bibles, the major picture sources of the Old Saxon Genesis. Freed from such a space constraint, the Old Saxon poem became amenable to extensive illustrations liberally distributed throughout the text, as was true of the Cotton Genesis originally and the Millstatt Genesis later on. This lifting of the space limitation imposed by the pandect format would have made possible a fuller execution of verse–picture collaboration in the Old Saxon Genesis as well, again prefiguring the Millstatt Genesis in this respect. Of particular interest in this connection is that the majority of images in Genesis B in Junius 11 were internally inspired (Table 15, section 12.1), and that they were integrated with their internal models into the elaborate system of pictorial organization (Tables 16 and 17, section 12.2). Their inherent alignment with the portions derived from the Old Saxon Genesis may make it more likely than not that these internally motivated illustrations were constructed by the original Carolingian artist, rather than added secondarily by an Anglo-Saxon copyist in the course of manuscript transmission. Had an insular artist been responsible instead, the image multiplication in question would have proceeded along with or following translation into Old English. Since the internal basis of image construction presupposes an active command of and full commitment to the living Touronian style—rather than superficial borrowings and mechanical replications of earlier images as suggested by Doane (1991: 41), this would mean that the Anglo-Saxon artist concerned would have been such a competent practitioner, capable of working with and elaborating on the systematicity of pictorial representation deeprooted in the Touronian tradition. However, given that, as reasoned below, the framing

300 

 12 The Touronian Foundations of the Old English Genesis: A Synthesis

of the continental material with the insular work took place toward the end of the ninth century—almost two generations after the demise of the underlying artisanship—the assumed Anglo-Saxon contribution, far removed in space and time from the original environment, would appear largely anachronistic. Under these circumstances completely detached from the original environment of the evolving Touronian tradition, outrageous deformations and reanalyses would have occurred on a larger scale and with a higher frequency than actually happened—realized in fact moderately and locally as the reconfiguration and decategorization of the book and the scepter (section 12.7.1) and the recomposition of the Temptation of Adam as that of Eve (section 12.7.2), and characterized as minor exceptions to the prevailing strict adherence to the Touronian heritage. The continental lineage of the Junius pictures thus secured may lead in turn to the inference that, upon removal of the space constraint pointed out above, the Carolingian artist needed more pictures than were available from his immediate sources, the Touronian Bibles. Accordingly, he would have taken the liberty of creating further images, drawing on the models of his own making earlier in the manuscript. Thus, the conspicuous proliferation of images with similar motifs, an apparently idiosyncratic feature of the Old English Genesis in Junius 11, may succumb to a principled account in terms of the specific context of transformation in its standing from part of a pandect to a book on its own (or a selection of episodes thereof). Subsequently, sometime during the second half of the ninth century, a copy of the illustrated Old Saxon Genesis would have been transmitted to Wessex, perhaps coupled with the Heliand (cf. Raw 1976: 148; Doane 2013: 51–54). Mediated through Wessex– West Francia alliance, the Anglo-Saxon acquisition of this illustrated Old Saxon Genesis manuscript may have occurred as early as 856 on the occasion of King Æthelwulf of Wessex’s marriage to Judith, Charles the Bald’s daughter, as conceived by Raw (1976: 148) and Doane (1991: 52–53). Charles’s patronage of the abbey of St. Martin at Tours, as most eloquently represented by the Vivian Bible, particularly its presentation miniature (423r), constitutes telling circumstantial evidence for the inference that the illuminated copy, originally produced in praise of the king,148 was bestowed on his son-in-law. Alternatively, the manuscript may have arrived much later, shortly before its compilation with Genesis A around the year 900. The latter route of transmission would have run via East Francia rather than directly from West Francia as with the first channel of transmission. The cultural connection between Tours and monasteries beyond the borders to the east, such as at Trier or Prüm (section 1.2), may lend plausibility to a direct descent from Old Saxony. More specifically, John the Old Saxon (fl. 890), abbot of Athelney by King Alfred’s appointment and collaborator in his project of revival of learning, may have been not only responsible for the introduction of the Old Saxon copy

148 It may be recalled that the frontispiece, Pii, God enthroned, in Junius 11 is specifically modeled on the portrait of Charles in the presentation scene (423r) in the Vivian Bible.

12.8 On contextualizing the production of an illustrated manuscript of the Old English Genesis 

 301

to England, but also involved, in whatever capacity, in the production of the unified Old English Genesis (Keynes/Lapidge 1983: 93 [Asser’s Life of King Alfred, Chap. 78], 103 [Asser’s Life of King Alfred, Chap. 94], 126 [Prose Preface to Gregory’s Pastoral Care], 260n169; Doane 2011: 67–69). No less plausibly, Grimbald of St. Bertin’s (d. 901), another advisor to Alfred’s cultural program, would have brought an Old Saxon Genesis copy to Wessex, along with other notable manuscripts, and would have been involved in its West-Saxonization (Keynes/Lapidge 1983: 93 [Asser’s Life of King Alfred, Chap. 78], 126 [Prose Preface to Gregory’s Pastoral Care], 214n26, 260n168; Doane 2011: 67–69). In any event, the Old Saxon Genesis brought from the continent was embedded into the Anglian Genesis (an antecedent of Genesis A; section 1.1; see also sections 1.1 and 2.1, and Chapter 5, Volume 2) in the last quarter of the ninth century or around 900 at the latest (cf. Doane 1991: 52; 2011: 65; 2013: 40). This terminus ante quem rests on the notable occurrence of early West-Saxon linguistic features in the extant Old English Genesis, particularly in Genesis Aa (= Sievers’s Genesis I, the earlier portion of Genesis A preceding Genesis B; section 1.1) and Genesis B, in contrast to Genesis Ab (= Sievers’s Genesis III, the later portion of Genesis A following Genesis B). Dating to the last quarter of the century, rather than the third, may be justified to some extent by the consolidation of the Anglo-Saxons and the concomitant stabilization of West-Saxon society at that time, after a successful defense against Viking raids.149 Such sociocultural conditions would have been favorable to implementation of a literary project such as is postulated. Further, it may be naturally conjectured that this textual and pictorial recomposition constituted part of the cultural and educational program pursued at the court of Alfred the Great (Doane 1991: 52). As a consequence, the first three picture groups—the Fall of the Rebel Angels cycle, the Creation cycle, and the Adam and Eve cycle—along with relevant portions of the original text, were incorporated, through borrowing, translation, and adaptation, into the Old English Genesis (the Anglian Genesis), as outlined in section 1.1 above. The intensive editing that was exercised concurrent with and/or subsequent to the translation would have resulted in the metrical convergence of Genesis A1 (1–234, 852–966) and Genesis B (235–851) in distinction from Genesis A2 (967–2936), whereby, as substantiated in section 1.1, Genesis A1 and Genesis B are commonly distinguished from Genesis A2 by their significantly greater likelihood of manifesting double alliteration.150 The metrical integrity of Genesis A1 and Genesis B in opposition to Genesis A2 is evidenced not only by their overall higher probability of double alliteration, as demon-

149 As epitomized among others by the treaty between Alfred and Guthrum, king of the Vikings in East Anglia, about 886 (Keynes/Lapidge 1983: 171–172), and symbolized by Asser’s reference to Alfred as Angul-Saxonum rex, presumably reflecting current use in the 880s and thereafter (Keynes/Lapidge 1983: 67 [Asser’s Life of King Alfred, Chap. 1], 227–228n1). 150 The following is a brief exposition of the metrical (re)organization of the Old English Genesis immediately bearing on the Anglo-Saxon reuse of the Touronian pictures. For fuller accounts, see, in particular, section 4.4, Volume 2.

302 

 12 The Touronian Foundations of the Old English Genesis: A Synthesis

strated in section 1.1, but also by the way in which the incidence of double alliteration varies per section, as represented in Table 22. Specifically at stake is to determine how the rate of double alliteration fluctuates as sections move forward along the time axis. It is a reasonable expectation that the rate of double (aa) to single alliteration (ax/xa) in the a-verse, or the occurrence of double alliteration per 100 a-verses, makes no significant difference across sections of the same poem. Put another way, other things being equal, double alliteration is as likely to occur in one section as in another, in statistical terms, while subject to a limited range of variation as a matter of contingency. That is, the parameter of sections should exert no appreciable effect on the likelihood of double alliteration. This is equal to saying technically that, by linear regression analysis (for details, see below), the straight line fitting the scatter plot for the two variables (section and occurrence of double alliteration per 100 a-verses) does not differ significantly from the response mean line, that is, the base horizontal line with a constant value for the parameter of double alliteration rate fixed at the mean. The independence of alliterative pattern from section progression is substantiated, in conformity with the above prediction, in Beowulf and the Heliand, the prime representative epics in Old English and Old Saxon, respectively. Much the same applies to Genesis B as well. In Figures 11, 12, and 13, scatter plots for Beowulf, the Heliand, and Genesis B, respectively, each point represents the number of double alliterations per 100 a-verses (DA, the vertical axis) in a given section (S, the horizontal axis). 65 60 55 50 DA 45 40 35 30 25

0

10

20 S

30

40

Figure 11. Bivariate analysis of sections and double alliteration in Beowulf

12.8 On contextualizing the production of an illustrated manuscript of the Old English Genesis 

85 80 75 70 DA 65 60 55 50 45

0

10

20

30

40 S

50

60

70

80

Figure 12. Bivariate analysis of sections and double alliteration in the Heliand

75 70 65 DA

60 55 50 45 40

4

5

6

7

8 S

9

10

11

Figure 13. Bivariate analysis of sections and double alliteration in Genesis B

12

 303

304 

 12 The Touronian Foundations of the Old English Genesis: A Synthesis

80 70 60 DA

50 40 30 20 10

0

5

10

15

S

20

25

30

35

Figure 14. Bivariate analysis of sections and double alliteration in Genesis A

By linear regression analysis,151 the relationship between these two variables may be expressed by means of best fit, a linear model of DA = a + bS (DA: double alliteration; S: section; a: intercept; b: estimated regression coefficient for ß, a corresponding unknown regression coefficient in the population, Beowulf, for example). As represented by the descending or ascending straight lines in the above figures, the mathematical equations postulated for the three poems in question are: DA = 51.028592 – 0.0821282S (Beowulf); DA = 63.134004 + 0.0033199S (Heliand); DA = 64.285714  – 0.6785714S (Genesis B). The probability of obtaining the given estimated regression coefficients (b ≠ 0; see the column under Estimate in Table 21) must then be computed against the null hypothesis of ß = 0, namely, that the population value of the regression coefficient ß on section is zero; put another way, the variable of sections is hypothesized by default to have no effect on the occurrence of double alliteration in the populations—Beowulf, Heliand, Genesis B, and Genesis A, respectively—from which the above distribution patterns (Figs. 11–14) are drawn as their respective samples. Accordingly, at stake is to determine whether to reject the null hypothesis that ß = 0 in favor of the alternate that ß ≠ 0 on the basis of how much b differs from ß, as technically measured by the t-ratio (tr; see the column under t-Ratio in Table 21): tr = b/sb (sb: the standard error of b; see the column under Std Error in Table 21).

151 JMP13.2.0, SAS Institute Inc, 2016. For more detailed accounts of simple linear regression, see, for example, Feinstein/Thomas (2002: 93–109, 149–180) and Schroeder/Sjoquist/Stephan (2017: 1–20, 31–52).

12.8 On contextualizing the production of an illustrated manuscript of the Old English Genesis 

 305

Here one need not be concerned with technical details of the computational basis involved. Of immediate interest, the standard error of b (sb) expresses the variability, among different samples, of regression coefficient estimates (b’s) for the same population parameter (ß). Simply put, the larger the standard error is, the less confident one may be of the true value of ß because of the greater variability of its estimates. Given that sb constitutes the denominator of the t-ratio (as above), a smaller value for sb, with its concomitant closer approximation to ß, leads to a larger t-ratio. Since the t-ratio follows the Student’s t-distribution, a viable statistic for statistical hypothesis testing, one can compute—under the assumption that the null hypothesis of ß = 0 is true—the probability of getting a t-ratio larger in absolute value (|t|) than the one actually obtained from the sample. If the computed probability value turns out to be less than 0.05 (p < 0.05; see the column under Prob > |t| in Table 21), the estimated regression coefficient b under consideration would seem unlikely to have been obtained only by chance, as far as the null-hypothesized zero value of ß is tenable. Therefore, in light of such an extremely low probability, one should instead reject the underlying null hypothesis that ß = 0. In other words, the hypothesized independence of the two variables must be discarded at the 0.05 level of significance in favor of the alternative conceptualization that they are functionally related; in the present context specifically, one may be justified in inferring that the forward movement of sections has a certain effect on the occurrence of double alliteration, now that the alternate hypothesis of ß ≠ 0 holds true. Conversely, insofar as the t-ratio in absolute value is small enough for there to be a probability higher than 0.05 that a larger value for the t-ratio may be available (that is, the calculated statistic does not fall in the critical region of 2.5 percent on either side of the t-distribution curve), one cannot reject the null hypothesis of ß = 0, meaning that sections and alliterative pattern should be regarded as totally independent variables. Since the p-values yielded from Beowulf, the Heliand, and Genesis B are each well above the 0.05 level of significance (Table 21), one cannot reject the null hypothesis: the estimated regression coefficients on the variable of sections for these poems are thus statistically no different from zero at the 0.05 level. Therefore, it must be concluded that section progression has no effect on implementation of alliterative pattern in the a-verse in the three poems. Table 21. Parameter estimates report for the regression coefficient on S in Beowulf, the Heliand, Genesis A, and Genesis B Poem

Estimate (b)

Std Error (sb)

t-Ratio

Prob > |t|

Beowulf Heliand Genesis A Genesis B

–0.082128  0.003319 –0.68695 –0.678571

2.038897 0.044645 0.159328 1.878381

–1.00  0.07 –4.31 –0.36

0.3229 0.9409 0.0002 0.7327

306 

 12 The Touronian Foundations of the Old English Genesis: A Synthesis

Table 22. Frequency of double alliteration per 100 a-verses according to sections in the Old English Genesis Section

Frequency of double alliteration per 100 a-verses

Genesis A/B

1  2  3  4  5  6  7  8  11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27.1 27.2 28 29 30 31 32 33 34 35 36 37 38 39 40 41

62 42 52 48 73 43 61 64 55 56 60 37 56 45 40 31 27 42 36 29 42 39 32 33 44 51 46 43 29 29 32 20 43 39 31 17 28 23 35 28

A  A  A  A  B  B  B  B  B  B  B  B+A A  A  A  A  A  A  A  A  A  A  A  A  A  A  A  A  A  A  A  A  A  A  A  A  A  A  A  A 

12.8 On contextualizing the production of an illustrated manuscript of the Old English Genesis 

 307

Such an absolute lack of dependence of alliterative pattern on section development fails to be borne out by Genesis A, when taken as a whole (Fig. 14 and Tables 21 and 22).152 By regression analysis, one may posit a linear model, DA = 48.647177 – 0.68695S, to account for the relationship between these two variables. Given a p-value of less than 0.01 obtained, it may be warranted to conclude that the proposed linear model with a regression coefficient (b = –0.68695) significantly differs from the response mean line in which the dependent variable of double alliteration rate keeps its average value (37.3125) constant in total disregard of the changing section value. In other words, whatever the exact causal relationship involved, the occurrence of double alliteration per 100 words in Genesis A apparently decreases as sections advance. This correlation is contrary to expectation, thereby calling for explanation. Radically different pictures present themselves when one examines comparable relations with respect to the set of Genesis A1 and Genesis B on the one hand (Fig. 15), and Genesis A2 on the other (Fig. 16). Since verse line 966, the end of Genesis A1 as defined in section 1.1 above, occurs in the middle of section 16, one may subsume this section in its entirety under Genesis A1 for the purpose of section-based comparisons here. Accordingly, Genesis A1/Genesis B ranges from section 1 through section 16 (where section 14 constitutes a combination of Genesis A and Genesis B), while Genesis A2 comprises sections 17 through 41. It should be noted that sections 9 and 10 are lacking in the extant manuscript (Table 22; see also Table 2, section 1.1). Inasmuch as one is concerned with the rate of double alliteration relative to single alliteration per section, neither the size of individual sections nor their status of preservation in the manuscript, complete or incomplete, seriously affects the investigation to be carried out below. As it turns out, the two linear models obtained (DA = 56.56044  – 0.3604396S for Genesis A1/Genesis B, and DA = 44.581197  – 0.3777778S for Genesis A2) do not differ significantly from the mean lines at the 0.05 level of significance (p = 0.606 for Genesis A1/Genesis B and p = 0.0832 for Genesis A2; Table 23). Accordingly, one cannot reject the null hypothesis that the occurrence of double alliteration is independent of the variable of sections. In other words, double alliteration is as likely to occur in one section as in another in either of the two Genesis portions, precisely the normal pattern as expected and in keeping with the other three poems. The statistically even distribution of double alliteration throughout the sections in the two parts of the Old English Genesis—Genesis A1/B and Genesis A2—implies that either portion is metrically self-contained on its own regarding implementation of alliteration in the a-verse, much as are Beowulf, the Heliand, and Genesis B. Put another way, rather than subsumed under a single entity, Genesis A1 and A2 are separate in terms of metrical organization, whereas the genealogically distinct segment, Genesis B, is integrated with Genesis A1. This split metrical identity of the originally single work, the

152 Section 14—a combination of Genesis B (821–851) and Genesis A (852–871)—is disregarded due to its metrical heterogeneity. All other sections are treated as continuous for the purpose of analysis.

308 

 12 The Touronian Foundations of the Old English Genesis: A Synthesis

Anglian Genesis (a preform of Genesis A), would have resulted through the interpolation of an Old English version of the Touronian Old Saxon Genesis episodes (an antecedent of Genesis B) and the ensuing revision of the whole work whereby the two adjacent constituents, Genesis A1 and Genesis B, are coordinated with each other and accordingly recomposed as a metrically single larger constituent (the Old English/Old Saxon Genesis). Because of the metrical adaptation to Genesis B with respect to the occurrence of double alliteration, Genesis A, taken as a whole (Fig. 14), appears to display a correlative pattern between operation of double alliteration and progression of sections. Simply put, double alliteration would seem less likely to occur in the sections as they are farther removed from the end of Genesis B. In the final analysis, however, this apparent correlation should not be characterized as an inherent, irreducible property of Genesis A; rather, it must be viewed merely as an epiphenomenon derived from the metrical integration with Genesis B of Genesis A1 in separation from Genesis A2, as elucidated above.

75 70 65 60 DA 55 50 45 40 35

0

2.5

5

7.5 S

10

12.5

15

Figure 15. Bivariate analysis of sections and double alliteration in Genesis A1 and Genesis B combined (sections 1–16)

12.8 On contextualizing the production of an illustrated manuscript of the Old English Genesis 

55 50 45 40 DA 35 30 25 20 15

10

15

20

25

S

30

35

40

45

Figure 16. Bivariate analysis of sections and double alliteration in Genesis A2 (sections 17–41)

60 55 50 45 DA

40 35 30 25 20 15

10

15

20

25

S

30

35

40

45

Figure 17. Bivariate analysis of sections and double alliteration in Genesis III (sections 15–41)

 309

310 

 12 The Touronian Foundations of the Old English Genesis: A Synthesis

Table 23. Parameter estimates report for the regression coefficient on S in Genesis A1/Genesis B, Genesis A2, and Genesis III Poem

Estimate (b)

Genesis A1/Genesis B (sections 1–16) –0.36044 Genesis A2 (sections 17–41) –0.377778 Genesis III (sections 15–41) –0.537493

Std Error (sb)

t-Ratio

Prob > |t|

0.680486 0.209008 0.192957

–0.53 –1.81 –2.79

0.6060 0.0832 0.0098

Of particular interest in this connection is Sievers’s (1929) bipartition of Genesis A before and after Genesis B, namely, Genesis I (verse lines 1–234) on the one hand, and Genesis III (verse lines 852–2936) on the other, as referred to in section 1.1 above. Sievers’s division would imply that, analogous to Genesis A2 in my conceptualization, Genesis III ought to be consistent and homogeneous in alliterative pattern insofar as it is characterized as a separate entity on its own, and accordingly display no significant relationship between double alliteration and section progression: more specifically, double alliteration would occur throughout this portion of the poem without any notable difference between its sections as they move forward.153 This prediction, however, proves to be empirically wrong (Fig. 17 and Table 23): with a p-value of 0.0098, the linear model for Genesis III (DA = 49.600712 – 0.5374932S) must be regarded as significantly different from the mean line with a zero regression coefficient. Therefore, one must conclude that double alliteration decreases in implementation in Genesis  III as sections proceed, an unexpected state of affairs beyond immediate understanding. Thus, Sievers’s view on the autonomy of Genesis III has not been verified. By alternatively subscribing to the division into Genesis A1 and A2 proposed above, the apparently increasing avoidance of double alliteration in correlation to a greater distance from the end of Genesis B must instead be rejected as epiphenomenal, lacking in empirical reality: this pseudo correlation between section and alliteration is simply an unsubstantial consequence of the incorrect assignment of verse lines 852 through 966 to the following—rather than the preceding—body of Genesis A, contrary to my contention. A question arises at this point: Why were specifically verse lines 1 through 234 and 852 through 966 (or, in terms of section divisions, sections 1–4 and 14–16) coordinated with Genesis B (sections 5–14) in their metrical organization? The answer to this challenging question was already suggested in section 1.1: This portion in the Old English Genesis constitutes a textual correspondent of the twenty-two pictures in the illustrated Old Saxon Genesis episodes from which the original text of Genesis B would have been extracted and interpolated into Genesis A. Covering the earliest stories in the

153 Before proceeding, one thing deserves particular notice. Given that section 14 comprises Genesis A and B, as remarked above, this heterogeneous section must be excluded from the section-based examination to be carried out. Accordingly, Genesis III ranges from section 15 through 41.

12.8 On contextualizing the production of an illustrated manuscript of the Old English Genesis 

 311

book of Genesis, the two Germanic biblical poems overlap considerably in their narrative content, if not in their manner of narration. Faced with the two related versions, one illustrated (Old Saxon), the other unillustrated (Old English), the Anglo-Saxon poet/ editor/scribe would have been intent on fruitfully synthesizing the two by their eclectic and complementary use. Because the intended audience/readership was primarily Anglo-Saxon, and because Genesis A (the Anglian Genesis) was obviously available with its far more extensive material coverage, the Old English text was naturally chosen as a basis of revision for a synthetic work. By contrast, in the absence of accompanying pictures in the Anglian Genesis manuscript at hand, all twenty-two images available were transposed from the Old Saxon Genesis episodes for reuse in the new text, although their original corresponding verses were not concurrently adopted in full into the synthesized version.154 Thus, all Touronian/Old Saxon pictures eventually found their way into the Old English Genesis in its initial parts, much as in the Old Saxon original. Accordingly, the incorporation of the Old Saxon material involved visual representations, which were in effect aligned with the verse passages up to Genesis A verse line 966 or sections 1 through 16. In this way, crossing the boundaries between the two distinct sources at 234/235 and 851/852 in the surviving manuscript, the text running from verse lines 1 through 966 was treated as a unitary whole because of its corresponding pictorial integrity, and as such subjected to the metrical integration leading to the enhanced implementation of double alliteration in opposition to Genesis A2 (967–2936 or sections 17–41). A further thinking on the unity of Genesis A1/Genesis B (verse lines 1 through 966) that I claim to have been grounded on the corresponding Old Saxon illustrated episodes may cast light on what appears to be inapposite text–image alignment in Junius 11, particularly characteristic of the first twenty-two pictures. Adducing this very property—the lack of optimal correspondence between text and picture in the extant Junius manuscript, specifically the alleged impropriety of picture spaces in terms of their size, number, and location—Broderick (1978: 474–475; 1983: 164–165; 2009: 385–387) raised objection to Raw’s derivation of Junius pictures from illustrations in an Old Saxon Genesis manuscript. Given the partial affinity of Raw’s view with mine as carefully delineated in section 1.3, Broderick’s refutation may be of immanent relevance here. According to his reasoning, the hypothetical models of an illustrated Old Saxon Genesis manuscript that Raw and I postulate (in different ways) would counterfactually have spared the Anglo-Saxon scribe and artist exactly the kind of errors that they actually

154 Apart from the narrative richness of the source material in the Old Saxon original, it remains unclear why exactly the passages corresponding to verse lines 235 through 851 in the extant manuscript were selected for interpolation. Specifically, one may wonder whether they were chosen for particular thematic reasons (Lucas 1992: 132–133; cf. Doane 2013: 30n78) or simply due to the loss or damage of the corresponding material in Genesis A (Doane 1991: 52n12, 54; 2013: 30–31). In any event, the material insufficiency in the available version of Genesis A should be invoked for the exceptional manipulation of the text.

312 

 12 The Touronian Foundations of the Old English Genesis: A Synthesis

made at their first experimentation in the absence of a direct model (in Broderick’s view). Correspondingly, the scribe and illustrator should have accomplished a far better text–image coordination in an Anglo-Saxon manuscript (Junius 11 or its antecedent) than they managed to execute in the extant form. In exercising such inference, Broderick seems to be driven by the conceptualization that the adoption/adaptation of the Old Saxon pictures for production of an illustrated Old English Genesis would largely have been a mechanical task, something similar to a literal copying of one and the same work. As argued from time to time (e.g., sections 1.1 and 12.8; and further Volume 2), however, the composition of the Old English Genesis was more a novel creation than a secondary reproduction. To say nothing of the translated and recomposed part to emerge eventually as Genesis B, the native Old English material (the Anglian Genesis resulting in Genesis A1) that fell under the picture cycles of the first twenty-two images would have hardly less differed in points of detail from the Old Saxon counterpart originally illustrated in the source manuscript. Thus, even in the face of a perfect model at hand for consultation in which pictures are properly aligned with the text (an assumption that is itself open to doubt), the alleged text–picture incongruence would inevitably have arisen in the earliest portion of the Anglo-Saxon Genesis (verse lines 1–966) in the making: transposed into a new Old English text yet to be adjusted and consolidated, the picture spaces would have had to be specified anew on an experimental basis. Thus, the pictorial spaces would have been inherently subject to alteration and variation in their size and alignment with the emergent text. In this light, the text–image disparity allegedly inherent in Junius 11 can hardly disprove the use of an illustrated Old Saxon Genesis manuscript as an exemplar.

13 Conclusion Synchronically defined as a textual body ranging from verse lines 1 through 966 of the Old English Genesis contained in MS Junius 11, the Old English/Old Saxon Genesis (Genesis A1 and Genesis B) is visualized by the first twenty-two pictures provided in the manuscript. The relative autonomy of this part is evidenced by its sharp metrical/pictorial-based demarcation from the rest of the Old English Genesis (Genesis A2). This relatively independent portion came into being as a result of interpolation of Genesis B— an Old English translation and redaction of Old Saxon Genesis episodes—into Genesis A (more precisely, the native Old English Genesis or the Anglian Genesis, anteceding Genesis A; section 2.1, Volume 2) and ensuing editorial adjustments. The original Old Saxon textuality underlying the extant Old English/Old Saxon Genesis thus subsists, partially and indirectly, as Genesis B, through translation, revision, and reorganization. Given the metrical and pictorial integrity of the Old English/Old Saxon Genesis as delineated in the existing manuscript and given the demonstrable Carolingian origin of the verse text, it may be hypothesized that the corresponding twenty-two pictures in Junius 11 also derived from the continent. More specifically reformulated in art historical terms, the Old English/Old Saxon Genesis would have been based on a manuscript of the Old Saxon Genesis (or more probably on a selection of Genesis episodes) illustrated with the three picture cycles reflected in the twenty-two Junius images—the Fall of the Rebel Angels cycle, the Creation cycle, and the Adam and Eve cycle. The reuse in the Old English/Old Saxon Genesis of the three picture cycles originally accompanying the Old Saxon Genesis narratives may thus be viewed as an exemplary case of miniatures “carried over into a new text recension” (Weitzmann 1947: 134–143). While the Old Saxon origin of the verse text is readily verified on philological grounds, the comparably continental pictorial foundations have been substantiated by intensive and extensive investigations carried out in this book. Exploring the Carolingian models and sources for the set of pictures thus delimited leads one to be primarily concerned with the original artist (also interchangeably and loosely referred to as the (pre-)Junius artist throughout this book; section 1.4) responsible for initially producing the Old Saxon Genesis illustrations, which are assumed to be reflected, with varying degrees of transparency, in the Anglo-Saxon Junius pictures at issue. Accordingly, of central interest in the foregoing explorations were, first, the cognitive foundations and artistic resources structurally and macroscopically conceived, that is, the inherited expert knowledge and craftmanship that the Carolingian illustrator drew on conventionally or creatively in designing novel pictures, and, second, the concrete acts of image productions dynamically and microscopically considered, that is, the array of operations, including adaptations, recompositions, and reorganizations, that he executed in interaction with diverse challenges posed by particular models and sources, visual and verbal. Constructed thus by the innovative artist on Carolingian soil, the twenty-two pictures corresponding to the Old English/Old Saxon Genesis are integrated into the system https://doi.org/10.1515/9783110788068-013

314 

 13 Conclusion

of pictorial organization, the multidimensional network of opposition, complementation, parallelism, and variation in conceptual and compositional terms. This system is predicated on the few formal parameters bearing on the ways in which God and other extraterrestrial beings are represented: (i) the types of figures (positive vs. negative, major vs. minor), (ii) the representation modes (detailed vs. reduced), and (iii) the types of scenes (state/event vs. action). Among all picture members, the pair of Pii/P2 lay the foundation for the system by constituting its core, serving as the compositional basis for all other pictures, and determining their positions within the interpictorial network. Far from a purely static/descriptive/classificatory generalization, the system of pictorial organization has a cognitive–constructive dimension. The system may be conceptualized as constituting part of the artist’s working knowledge, which he would have drawn on in his artistic practice. That is, largely motivated by metaphor and metonymy, and exercising analogical extension, the artist would have constructed novel images out of the inherited inventory of resources in accordance with the underlying system. Thus, a subset of the Junius pictures would have been motivated on purely internal grounds without corresponding models elsewhere. In short, the pictorial system would have had a cognitive reality in the artist’s mind as an overarching generative scheme for images in the making. Whereas in some cases new pictures were created primarily on the basis of internal models—the actualized and/or virtual images organized by the underlying pictorial system of the Touronian Old Saxon Genesis anteceding the Old English/Old Saxon Genesis—through the mechanism outlined in the preceding paragraph, in other cases, one may explore external models, that is, sources outside the Old English/Old Saxon Genesis picture cycles. The primary external models can be divided into three classes, according to differing combinations of two binary source features, Touronian/Non-Touronian (derived or not from the Touronian Bibles) and Cottonian/Non-Cottonian (derived or not from the Cotton Genesis family). As it turns out, the pictures of the Old English/Old Saxon Genesis originated, for the most part, from the Touronian Bibles and, more precisely, from the Touronian/Cottonian class in their majority and the Touronian/Non-Cottonian class in a few but privileged instances involving the cardinal pair Pii and P2. While the intersection of the Touronian and the Cottonian sets outnumbers, to a comparably overwhelming extent, the Touronian/Non-Cottonian and the Non-Touronian/Cottonian classes (Table 15, section 12.1), respectively, these two minor classes—in terms of membership size—are by no means of an equally peripheral status. In actuality, the Touronian identity—whether additionally Cottonian or not—is of essential significance. The primacy of Touronian membership is persuasively confirmed by the asymmetry in the ways in which the two respective sources contributed, relative to each other, to the composition of Junius pictures. Specifically, the Touronian/Non-Cottonian sources served as exclusive models for several Junius images (Pii, P2, and P3; Table 15, section 12.1). Of particular interest here, the prime members of the Junius pictorial system—Pii and P2—are thus founded exclusively on the Touronian/Non-Cottonian models, as remarked above. By diametrical contrast, the Creation cycle pictures (P6 and

13 Conclusion 

 315

P7)—the only images largely derived from Non-Touronian/Cottonian sources—were concurrently inspired by Touronian/Non-Cottonian representations (Table 15, section 12.1). Thus, capable of determining composition on its own, the Touronian identity prevailed over the Cottonian heritage, which depended on additional support from other sources. No less significant in weighing the Touronian and Cottonian identities is the opposition between the Touronian and Non-Touronian members of the Cotton Genesis family. At stake is the competition for shaping the composition of P31.2 between the two Cottonian Covering models—Bamberg 7v2 (Touronian) on the one hand and San Marco (Non-Touronian) on the other. Of paramount interest, it was the Touronian model that was eventually chosen, despite its deviation from the Cotton Genesis archetype, as preserved more faithfully at San Marco. In contrast to these three fully substantiated classes of external models, there are no demonstrable examples of the remaining logical possibility: Non-Touronian/Non-Cottonian external sources. Neither the Utrecht Psalter—the most likely candidate for this class—nor the Beatus Commentaries (see below) can be identified as primary sources for the Old English/Old Saxon Genesis pictures. The low profile of the Utrecht Psalter seems of particular interest, given its cultural, geographical, and temporal vicinity to the School of Tours. While Utrecht undoubtedly inspired a small number of Touronian Old Saxon Genesis pictures, its scope and sphere of influence must be regarded as limited and superficial. Accordingly, one may postulate a three-step scale of inspiration from external sources as follows: Touronian (regardless of Cottonian identities, represented by the Touronian Bibles) > Non-Touronian/Cottonian (represented by the San Marco mosaics) > Non-Touronian/Non-Cottonian (represented by the Utrecht Psalter). This ranking corresponds to the artist’s familiarity with and depth of knowledge of the external material available. Specifically, the artist’s working knowledge centered on the Touronian Bibles. The portion of the Cotton Genesis that is not reflected in the Touronian Bibles was of secondary importance, and still less prominent was the Utrecht Psalter. The Touronian foundations thus established beyond reasonable doubt concern not only the composition of individual images on the surface but, more significantly, the structural underpinnings of the pictorial system itself. Although still in its nascence, the generalization based on the binary features—state versus action and detailed versus reduced—pertaining to the representation of the ruler was latently available in the Touronian Bibles, embodied in the group of pictures that includes Vivian 423r. Thus, seizing on this generative potential, the artist would have brought it to full realization through his creative engagement in illuminating the Old Saxon Genesis episodes. In light of the scale of inspiration and the corresponding familiarity ranking posited above, and more importantly, given the nature of his artwork, the artist must be viewed as an active practitioner of the School of Tours. As such, the artist made full use of the rich resources accessible in the Touronian tradition, whether already exemplified or yet to be experimented. Since, naturally enough, the Touronian Bibles occupied the core of his working knowledge, he made exhaustive use of them in accordance with his

316 

 13 Conclusion

artistic interests and purposes. The artist’s reuse of the Touronian material involved an extremely broad range of pictorial composition, from individual figures and their fine attributes to their configuration to their narrative progression. Yet, far from copying the Touronian models mechanically or reproducing them faithfully, the artist imaginatively executed radical reinterpretation, extensive recomposition, and thorough reorganization, largely motivated by metaphor and metonymy, and guided by the pictorial system of his own making. The left-to-right narrative progression in the Touronian originals was reversed systematically to the right-to-left progression in the Old Saxon adaptation of Genesis. This thorough reversal in narrative directionality would likely have been determined by the consistent positioning of God on the right side of a pictorial zone and the concomitant leftward orientation of the God figure in Junius. Such alignment of God depicted in profile on the right side of the page presumably would have been profoundly indebted to the figuration of the Creator in P2, one of the prime paired images derived from Vivian 423r. In reworking the Touronian images, the artist often drew on multiple works and unified thematically varied material into a single composite scene. Occasionally, he went so far as to appropriate the models in their mirror image to accommodate them in conformity with the overall design of his new work. The Junius artist was not simply a consumer of the conventional sources available from the Cotton Genesis tradition in general and its Touronian manifestations in particular. On the contrary, he made notable contributions as an innovator of the tradition by supplying novel images to be used for subsequent family members. At issue is the depiction of the Judgment scenes (P41.1/P44), the scene type that had barely been established as an autonomous pictorial episode so far. After these pictures were constructed through the conceptual conversion of the conventional Reproval/Denial scene into an innovative image of the Judgment, they made their way outside the Old Saxon Genesis picture series—they were merged, through recomposition, into a single image in the San Marco mosaics. Of still greater import, not only in cross-cultural terms but also in immanent terms to the School of Tours itself, was the influence that P2—the profile member of the pair of pivotal images—exerted over a group of the Beatus Commentaries in northern Spain. The scene of God enthroned at the right, flanked by seraphim and facing a small child (Lucifer) standing in the center, gave inspiration to the Branch IIa of the Beatus Commentaries, along with other Touronian artistic features. This specific reflection of the Carolingian–Asturian cultural contact eloquently testifies that the Junius innovation was acknowledged as part of the Touronian tradition and was transferred beyond Carolingian soil in unity with other established properties of the School. Thus, the Old Saxon Genesis illustrations were not only Touronian offspring in their genealogical identities, but also robust constituents of the tradition capable of contributing to its enrichment. They are far from being an evolutionary dead end. While deeply rooted in the Touronian tradition in their emergence, the Old Saxon Genesis pictures were subject to alteration and reinterpretation in their subsequent

13 Conclusion 

 317

history of transmission as they were brought over to Anglo-Saxon England. The holding on the left lap of the book (below) and the scepter (above)—the complex configuration that was shaped by the double origins of Pii/P2, namely, Vivian 423r and the Touronian exemplar of Christ in Majesty—suffered obscuration largely owing to the unnatural pose involved, so much so that the two constituent objects were merged into one, resulting in something unique that defies definite categorization, that is, an enigmatic object, called here a scroll/scepter for lack of a better name. The Temptation of Adam was reinterpreted as the Temptation of Eve in line with the more familiar biblical tradition, once the Old Saxon story was relocated to Anglo-Saxon soil, where the underlying Old Saxon episode may not have been as well known in detail as on the continent. Since the artist’s concern was to provide illustrations of the Old Saxon Genesis episodes ending in the Expulsion, he had no opportunity to use the Burial of Abel, available in Bamberg 7v4, although he took full advantage of the remaining pictures contained in that source. Because the story of Cain and Abel in Genesis A did not happen to articulate with the Old Saxon Genesis through the interpolation that affected its earlier portion, there was no chance of this narrative’s visualization being foundationally inspired by the Touronian Bibles through the intermediary of the illustrated Old Saxon models, as with the preceding episodes. Thus, on the basis of the illustrated Old Saxon Genesis episodes, the Old English/Old Saxon Genesis—defined by the first three picture cycles ending in P46 and the corresponding text from verse lines 1 through 966 in Junius 11— was shaped, through intensive adaptation and reworking, as a relatively discrete component of the Old English Genesis, in distinction from its remaining body (Genesis A2), which was retained closer to the underlying Anglian Genesis—to be substantiated and consolidated in terms of metrical adaptation and integration in Volume 2, especially in Chapters 4 and 5.

Bibliography Primary sources Ælfric’s Old English Hexateuch: London, British Library, Cotton MS Claudius B IV. http://www.bl.uk/ manuscripts/FullDisplay.aspx?ref=Cotton_MS_Claudius_B_IV; last access April 29, 2022. Apotheosis of Boethius: Paris, Bibliothèque Nationale, Ms. lat. 6401. https://gallica.bnf.fr/ark:/12148/ btv1b105096135; last access April 29, 2022. Bamberg Bible (Alcuin Bible): Bamberg, Staattsbibliothek, Msc. Bibl. 1. http://digital.bib-bvb.de/view/ bvb_mets/viewer.0.6.5.jsp?folder_id=0&dvs=1651020665689~35&pid=14322344&locale=ja&usePid1=tr ue&usePid2=true; last access April 27, 2022. Bamberg Commentaries (Isaias Glossatus): Bamberg, Staatsbibliothek, Msc. Bibl. 76 (Bamberg Commentaries). https://www.bavarikon.de/object/bav:SBB-KHB-00000SBB00000047; last access April 29, 2022. Bamberg Daniel Commentary: Bamberg, Staatsbibliothek, Msc. Bibl. 22. Beato de la Universidad de Valladolid (2000–2002): El Beato de la Universidad de Valladolid: Original Conservado en la Biblioteca de Santa Cruz de la Universidad de Valladolid. 2 vols. Madrid: Testimonio Compañía Editorial. Beato de San Millán (2002): El Beato de San Millán, Códice 33: Original Conservado en la Biblioteca de la Real Academia de la Historia. Madrid: Testimonio Compañía Editorial. Beato del Monasterio (2004): Beato del Monasterio de Santa Maria la Real de Huelgas de Burgos. 2 vols. Valencia: Scriptorium. Behaghel, Otto/Taeger, Burkhard (eds.) (1996): Heliand und Genesis. 10th edn. Tübingen: Niemeyer. Bibel von Moutier-Grandval (1971): Die Bibel von Moutier-Grandval: British Museum Add. Ms. 10546. Bern: Verein Schweizerischer Lithographiebesitzer. Biblia sacra (1993): Biblia sacra: Codex membranaceus saeculi IX. Rome: Istituto poligrafico e Zecca dello Stato. Bodleian Library MS. Junius 11: Oxford, Bodleian Libraries, University of Oxford. https://digital.bodleian. ox.ac.uk/objects/d5e3a9fc-abaa-4649-ae48-be207ce8da15; last access June 18, 2022. Casanovas, Jaime Marqués/Dubler, Cesar E./Neuss, Wilhelm (eds.) (1962): Sancti Beati a Liebana in Apocalypsin Codex Gerundensis. 2 vols. Oltun: Graf. Codex Aureus of St. Emmerams: Munich, Staatsbibliothek, Clm 14000. https://www.digitale-sammlungen. de/en/view/bsb00057171?page=,1; last access April 29, 2022. Codex Urgellensis (1997–2002): Beatus de Liébana Codex Urgellensis. 2 vols. Madrid: Testimonio Compañía Editorial. Códice de Fernando I (1994): El Beato de Liébana: Códice de Fernando I y Doña Sancha cuyo Original se Encuentra en la Biblioteca Nacional, Vitr. 14-2. Barcelona: Moleiro. Códice de Manchester (2001): Beato de Liébana: Codice de Manchester. Valencia: Patrimonio Ediciones. Códice de Saint-Sever (1984): Comentarios al Apocalipsis y al Libro de Daniel: Edición Facsímil del Códice de la Abadía de Saint-Sever Conservado en la Biblioteca Nacional de Paris bajo la Signatura Ms. lat. 8878 = Commentaires sur l’Apocalypse et le Livre de Daniel: Édition en fac-similé du manuscrit de l’abbaye de Saint-Sever conservé à la Bibliothèque nationale de Paris sous la Cote Ms. lat. 8878. 2 vols. Madrid: Edilán. Códice de Silos (2001–2003): Códice de Santo Domingo de Silos. 2 vols. Barcelona: Moleiro. Coogan, Michael D. (ed.) (2010): The new Oxford annotated Bible. New Revised Standard Version with the Apocrypha. An ecumenical study Bible. 4th edn. Oxford: Oxford University Press. De laudibus sanctae crucis: Vienna, Österreichische Nationalbibliothek, Cod. 652, fol. 2v. https://onb. digital/result/1127947F; last access April 29, 2022. https://doi.org/10.1515/9783110788068-014

320 

 Bibliography

Doane, A. N. (ed.) (1991): The Saxon Genesis: An edition of the West Saxon Genesis B and the Old Saxon Vatican Genesis. Madison, WI: University of Wisconsin Press. Doane, A. N. (ed.) (2013): Genesis A: A new edition, revised. Tempe, AZ: Arizona Center for Medieval and Renaissance Studies. Dufay Gospels: Paris, Bibliothèque Nationale, Ms. lat. 9385. https://gallica.bnf.fr/ark:/12148/ btv1b8426037h; last access April 29, 2022. Ebbo Gospels: Épernay, Bibliothèque Municipale, Ms. 1. https://portail.biblissima.fr/ark:/43093/mdata95f62 2eb35512469f5316beae13f0542ce0f0f4d; last access April 29, 2022. Egerton Genesis: London, British Library, Egerton MS 1894. http://www.bl.uk/manuscripts/FullDisplay. aspx?ref=Egerton_MS_1894; last access April 29, 2022. Epistulae Pauli: Munich, Bayerische Staatsbibliothek, Clm 14345. https://www.digitale-sammlungen.de/en/ view/bsb00024480?page=1; last access April 29, 2022. Escorial Beatus: Escorial, Biblioteca del Monasterio, &. II.5. https://rbdigital.realbiblioteca.es/s/rbme/ item/13137#?c=&m=&s=&cv=&xywh=-2449%2C-208%2C7392%2C4159; last access April 28, 2022. Facundus Beatus: Madrid, Biblioteca Nacional, MS Vitrina 14-2. http://bdh.bne.es/bnesearch/ detalle/1806167; last access April 28, 2022. Fulk, R. D./Bjork, Robert E./Niles, John D. (eds.) (2008): Klaeber’s Beowulf and the Fight at Finnsburg, 4th edn. Based on the third edition of Beowulf and the Fight at Finnsburg, ed. by Fr. Klaeber (1950). Toronto: University of Toronto Press. Gollancz, Sir Israel (ed.) (1927): The Caedmon manuscript of Anglo-Saxon Biblical poetry, Junius XI in the Bodleian Library. Oxford: Oxford University Press. Gospels of St. Médard de Soissons: Paris, Bibliothèque Nationale, Ms. lat. 8850. https://gallica.bnf.fr/ ark:/12148/btv1b8452550p; last access April 29, 2022. Grandval Bible (Moutier-Grandval Bible): London, British Library, Add MS 10546. http://www.bl.uk/ manuscripts/Viewer.aspx?ref=add_ms_10546; last access April 27, 2022. Green, Rosalie/Evans, Michael/Bischoff, Christine/Curschmann, Michael (eds.) (1979): Hortus Deliciarum. Vol. 1: Commentary. Vol. 2: Reconstruction. Leiden: Brill. Hamano, Akihiro (2016): Die frühmittelhochdeutsche Genesis: Synoptische Ausgabe nach der Wiener, Millstätter und Vorauer Handschrift. Berlin: De Gruyter. Harley Golden Gospels: London, British Library, Harley MS 2788. http://www.bl.uk/manuscripts/FullDisplay. aspx?ref=harley_ms_2788; last access April 29, 2022. Hildesheim Bronze Door of St. Michael’s Church: https://en.wikipedia.org/wiki/Bernward_Doors#/media/ File:Bernwardstür.jpg; last access April 27, 2022. Histoire Universelle en Prose: Vienna, Österreichische Nationalbibliothek, Cod. 2576. https://digital.onb. ac.at/RepViewer/viewer.faces?doc=DTL_3452052&order=1&view=SINGLE; last access April 29, 2022. Huelgas Beatus: New York, Pierpont Morgan Library, MS M.429. https://www.themorgan.org/collection/ Las-Huelgas-Apocalypse/1; last access April 30, 2022. Keynes, Simon/Lapidge, Michael (1983): Alfred the Great: Asser’s Life of King Alfred and other contemporary sources. London: Penguin. Klaeber, Fr. (ed.) (1913): The later Genesis and other Old English and Old Saxon texts relating to the Fall of Man. Heidelberg: Winter. Krapp, George Philip (ed.) (1931): The Junius manuscript. New York: Columbia University Press. Le Mans Gospels: Paris, Bibliothèque Nationale, Ms. lat. 261. https://gallica.bnf.fr/ark:/12148/ btv1b8427443x/f1.item; last access April 29, 2022. Lothair Gospels: Paris, Bibliothèque Nationale, Ms. lat. 266. https://gallica.bnf.fr/ark:/12148/btv1b8451637v; last access April 29, 2022. Millstatt Genesis: Klagenfurt, Kärntner Landesarchiv, Cod. GV 6/19. https://landesarchiv.ktn.gv.at/klais/at/ jr/iis/imdas/web/loadMask/view-mask-felder.jsf?objectId=2727409&maskId=null&maskName=null; last access April 27, 2022.

Primary sources 

 321

Millstätter Genesis (1967): Millstätter Genesis und Physiologus Handschrift. Vollständige Facsimileausgabe der Sammelhandschrift 6/19 des Geschichtsvereines für Kärnten im Kärntner Landesarchiv, Klagenfurt. 2 vols. Graz: Akademische Druck- u. Verlagsanstalt. Morgan Beatus: New York, Pierpont Morgan Library, MS M.644. https://www.themorgan.org/ manuscript/110807; last access April 28, 2022. Octateuch: Vatican, Biblioteca Apostolica, Vat. gr. 747. https://digi.vatlib.it/mss/detail/Vat.gr.747; last access April 29, 2022. Old Saxon Genesis fragments: Vatican, Biblioteca Apostolica, Palatinus Latinus 1447. https://digi.vatlib.it/ view/MSS_Pal.lat.1447; last access April 28, 2022. Osma Beatus: Burgo de Osma, Archivo de la Catedral, Cod. 1. https://arachne.dainst.org/entity/3175554; last access April 28, 2022. Psalter of Charles the Bald: Paris, Bibliothèque Nationale, Ms. lat. 1152. https://gallica.bnf.fr/ark:/12148/ btv1b55001423q; last access April 29, 2022. Roda Bible: Paris, Bibliothèque Nationale, Ms. lat. 6 (3). https://gallica.bnf.fr/ark:/12148/btv1b90658394; last access April 27, 2022. Rylands Beatus: Manchester, John Rylands University Library, Latin MS 8. https://www.digitalcollections. manchester.ac.uk/view/MS-LATIN-00008/1; last access April 28, 2022. St. Aure Gospels: Paris, Bibliothèque de L’Arsenal Ms. 1171. https://gallica.bnf.fr/ark:/12148/btv1b55000807v; last access April 29, 2022. Saint-Sever Beatus: Paris, Bibliothèque Nationale, Ms. lat. 8878. https://gallica.bnf.fr/ark:/12148/ btv1b52505441p; last access April 28, 2022. San Marco mosaics: Venice, Basilica di San Marco. Interior. Mosaics. Atrium. West Arm. South Cupola (The Creation Cupola). Dumbarton Oaks Research Library and Collection. https://images.hollis.harvard. edu/primo-explore/search?query=any,contains,san%20marco&tab=default_tab&search_scope=HVD_ DDO_IMG&vid=HVD_IMAGES&lang=en_US&offset=0; last access April 27, 2022. San Millán Beatus: Madrid, Real Academia de la Historia, Cod. 33. https://bibliotecadigital.rah.es/es/ consulta/registro.do?id=124; last access April 29, 2022. San Paolo fuori le Mura frescoes: Vatican, Biblioteca Apostolica, Cod. Barb. lat. 4406. https://digi.vatlib.it/ view/MSS_Barb.lat.4406; last access April 28, 2022. Schäfer, Frank (2019): Die Millstätter Genesis: Edition und Studien zur Überlieferung. Teil 1: Einführung und Text. Teil 2: Untersuchungen. Göttingen: Vandenhoeck. Silos Beatus: London, British Library, Add. MS. 11695. http://www.bl.uk/manuscripts/FullDisplay. aspx?ref=Add_MS_11695; last access April 28, 2022. Utrecht Psalter: Utrecht, Universiteitsbibliotheek, MS Bibl. Rhenotraiectinae I Nr 32. https://psalter.library. uu.nl/page/1; last access April 27, 2022. Utrecht Psalter (1984): Vollständige Faksimile-Ausgabe im Originalformat der Handschrift 32, Utrecht-Psalter, aus dem Besitz der Bibliotheek der Rijksuniversiteit te Utrecht. 2 vols. Graz: Akademische Druck- u. Verlags­ anstalt. Utrecht Psalter (2015): Der Utrecht Psalter: Utrecht, Bibliotheek der Rijksuniversiteit, Hs. 32. Kommentar/ commentary Koert van der Horst. Graz: Akademische Druck- u. Verlagsanstalt. Valladolid Beatus: Valladolid, Biblioteca de la Universidad, MS 433. https://almena.uva.es/discovery/ fulldisplay?docid=alma991000080639705774&context=L&vid=34BUC_UVA:VU1&lang=es&search_ scope=FISICO_ELECTRO&adaptor=Local%20Search%20Engine&tab=LibraryCatalog&query=any,con tains,ms%20433&facet=topic,include,Biblia.%20N.T.%20Apocalipsis%20Comentarios&offset=0; last access April 28, 2022. Vivian Bible (First Bible of Charles the Bald): Paris, Bibliothèque Nationale, MS lat. 1. https://gallica.bnf.fr/ ark:/12148/btv1b8455903b; last access April 27, 2022. Waetzoldt, Stephan (1964): Die Kopien des 17. Jahrhunderts nach Mosaiken und Wandmalereien in Rom. Vienna: Schroll.

322 

 Bibliography

Weitzmann, Kurt/Kessler, Herbert L. (1986): The Cotton Genesis: British Library Codex Cotton Otho B. VI. Princeton, NJ: Princeton University Press. Williams, John/Shailor, Barbara A. (1991): A Spanish apocalypse: The Morgan Beatus manuscript. New York: Braziller.

Secondary sources Alexander, J. J. G. (1978): Insular manuscripts, 6th to the 9th century. London: Miller. Anlezark, Daniel (2011): Old Testament narratives. Cambridge, MA: Harvard University Press. Austin, J. L. (1962): How to do things with words. London: Clarendon. Bergman, Robert P. (1980): The Salerno ivories: Ars sacra from medieval Amalfi. Cambridge, MA: Harvard University Press. Blum, Pamela Z. (1976): The cryptic creation cycle in Ms. Junius xi. In: Gesta 15, 211–226. Broderick, Herbert R. (1978): The iconographic and compositional sources of the drawings in Oxford, Bodleian Library, MS Junius 11. Ph.D. dissertation, Columbia University. Broderick, Herbert R. (1982): Some attitudes toward the frame in Anglo-Saxon manuscripts of the tenth and eleventh centuries. In: Artibus et Historiae 5, 31–42. Broderick, Herbert R. (1983): Observations on the method of illustration in MS Junius 11 and the relationship of the drawings to the text. In: Scriptorium 37, 161–177. Broderick, Herbert R. (2009): Metatextuality, sexuality and intervisuality in MS Junius 11. In: Word & Image: A Journal of Verbal/Visual Enquiry 25, 384–401. Broderick, Herbert R. (2014): Northern light: Late Anglo-Saxon Genesis illustration and the mosaics of San Marco. In: Büchsel/Kessler/Müller (eds.), 211–229. Büchsel, Martin/Kessler, Herbert L./Müller, Rebecca (eds.) (2014): The atrium of San Marco in Venice: The Genesis and medieval reality of the Genesis mosaics. Berlin: Gebr. Mann. Cardinali, Marco (ed.) (2009): La bibbia carolingia dell’Abbazia di San Paolo fuori le Mura. Città del Vaticano: Edizioni Abbazia San Paolo. Cook, Robin (2007): The Genesis page of the Alcuin Bible: Its layout and visual repertoire. In: Journal of the Australian Early Medieval Association 3, 113–129. Dell’Acqua, Francesca/Cutler, Anthony/Kessler, Herbert L./Shalem, Avinoam/Wolf, Gerhard (eds.) (2016): The Salerno ivories: Objects, histories, contexts. Berlin: Gebr. Mann. Demus, Otto (1984): The mosaics of San Marco in Venice. Vol. 2: The thirteenth century. Part 1: Text. Part 2: Plates. Chicago: University of Chicago Press. DeWald, Ernest T. (1932): The illustrations of the Utrecht Psalter. Princeton, NJ: Princeton University Press. Diebold, William J. (1990): The artistic patronage of Charles the Bald. Ph.D. dissertation, The Johns Hopkins University. Doane, A. N. (2011): The transmission of Genesis B. In: Sauer, Hans/Story, Joanna (eds.), Anglo-Saxon England and the Continent, 63–81. Tempe, AZ: Arizona Center for Medieval and Renaissance Studies. Dorfman, Eugene (1969): The narreme in the medieval romance epic: An introduction to narrative structures. Toronto: University of Toronto Press. Dufrenne, Suzy (1978): Les illustrations du Psautier d’Utrecht: Sources et apport carolingien. Paris: Ophrys. Dutton, Paul Edward/Kessler, Herbert L. (1997): The poetry and paintings of the First Bible of Charles the Bald. Ann Arbor, MI: University of Michigan Press. Ericksen, Janet Schrunk (2001): Offering the forbidden fruit in MS. Junius 11. In: Davidson, Clifford (ed.), Gesture in medieval drama and art, 48–65. Kalamazoo, MI: Medieval Institute Publications, Western Michigan University.

Secondary sources 

 323

Feinstein, Charles H./Thomas, Mark (2002): Making history count: A primer in quantitative methods for historians. Cambridge: Cambridge University Press. Finnegan, Robert (1998): The man in ‘nowhere’: A previously undiscovered drawing in Bodleian MS. Junius 11. In: English Studies 79, 23–32. Fischer, Bonifatius (1957): Die Alkuin-Bibel. Freiburg: Herder. Gaehde, Joachim E. (1963): The painters of the Carolingian Bible manuscript of San Paolo fuori le Mura in Rome. Ph.D. dissertation, New York University. Gaehde, Joachim E. (1971): The Touronian sources of the Bible of San Paolo fuori le Mura in Rome. In: Frühmittelalteriche Studien 5, 359–400. Gameson, Richard (1995): The role of art in the late Anglo-Saxon church. Oxford: Clarendon. Ganz, Peter (1978): Eduard Sievers. In: Beiträge zur Geschichte der deutschen Sprache und Literatur 100, 40–85. Garipzanov, Ildar H. (2008): The symbolic language of royal authority in the Carolingian world (c. 751–877). Leiden: Brill. Gneuss, Helmut/Lapidge, Michael (2014): Anglo-Saxon manuscripts: A bibliographical handlist of manuscripts and manuscript fragments written or owned in England up to 1100. Toronto: University of Toronto Press. Green, Rosalie B. (1955): The Adam and Eve cycle in the Hortus Deliciarum. In: Weitzmann, Kurt (ed.), Late classical and mediaeval studies in honor of Albert Mathias Friend, Jr., 340–347. Princeton, NJ: Princeton University Press. Grice, Herbert Paul (1975): Logic and conversation. In: Cole, Peter/Morgan, Jerry (eds.), Syntax and semantics 3: Speech acts, 41–58. New York: Academic Press. Henderson, George (1962): Late-Antique influences in some English mediaeval illustrations of Genesis. In: Journal of the Warburg and Courtauld Institutes 25, 172–198. Henderson, George (1975): The programme of illustrations in Bodleian MS Junius XI. In: Robertson, Giles/ Henderson, George (eds.), Studies in memory of David Talbot Rice, 113–145. Edinburgh: Edinburgh University Press. Hilmo, Maidie (2004): Medieval images, icons and illustrated English literary texts: From Ruthwell Cross to the Ellesmere Chaucer. Aldershot: Ashgate. Horst, Koert van der (1996): The Utrecht Psalter: Picturing the psalms of David. In: Horst, Koert van der/ Noel, William/Wüstefeld, Wilhelmina C. M. (eds.), The Utrecht Psalter in medieval art: Picturing the psalms of David, 22–84. Tuurdijk: HES. Hubert, Jean/Porcher, Jean/Volbach, W. F. (1970): Carolingian art. London: Thames. [English translation of L’Empire carolingien, Paris: Gallimard, 1968.] Jolly, Penny Howell (1997): Made in God’s image? Eve and Adam in the Genesis mosaics at San Marco, Venice. Berkeley: University of California Press. Joslin, Mary Coker/Watson, Carolyn Coker Joslin (2001): The Egerton Genesis. London: British Library. Karkov, Catherine E. (2001): Text and picture in Anglo-Saxon England: Narrative strategies in the Junius 11 manuscript. Cambridge: Cambridge University Press. Karkov, Catherine E. (2007): Margins and marginalization: Representations of Eve. In: Keefer, Sarah Larratt/ Bremmer, Rolf H. (eds.), Signs on the edge: Space, text and margin in medieval manuscripts, 57–84. Paris: Peeters. Karkov, Catherine E. (2011): The art of Anglo-Saxon England. Woodbridge: Boydell. Kaspersen, Søren (1988): Cotton-Genesis, die Toursbibeln und die Bronzetüren: Vorlage und Aktualität. In: Bernwardinische Kunst. Bericht über ein wissenschaftliches Symposium in Hildesheim von 10. 10 bis 13. 10 1984, 79–103. Göttingen: Goltze. Ker, N. R. (1957): Catalogue of manuscripts containing Anglo-Saxon. Oxford: Clarendon. Kessler, Herbert L. (1966): An eleventh-century ivory plaque from south Italy and the Cassinese revival. In: Jahrbuch der Berliner Museen 8, 67–95. Kessler, Herbert L. (1977): The illustrated Bibles from Tours. Princeton, NJ: Princeton University Press.

324 

 Bibliography

Kessler, Herbert L. (2009): Memory and models: The interplay of patterns and practice in the mosaics of San Marco in Venice. In: Medioeve: Immagine e memoria. Atti del Convegno internazionale di studi, Parma, 23–28 settembre 2008, 463–475. Milan: Electa. Kessler, Herbert L. (2009–2010): The Cotton Genesis and creation in the San Marco mosaics. In: Cahiers Archéologiques 53, 17–32. Kessler, Herbert L. (2013): Artistic reciprocity between Venice and Salerno in the thirteenth century. In: De Giorgi, Manuela/Hoffmann, Annette/Suthor, Nicola (eds.), Synergies in visual culture. Bildkulturen im Dialog. Festschrift für Gerhard Wolf, 407–419. Munich: Fink. Kessler, Herbert L. (2014a): Introduction. In: Büchsel/Kessler/Müller (eds.), 9–18. Kessler, Herbert L. (2014b): Thirteenth-century Venetian revisions of the Cotton Genesis cycle. In: Büchsel/ Kessler/Müller (eds.), 75–94. Klein, Peter K. (1976): Der ältere Beatus-Kodex Vitr. 14–1 der Biblioteca Nacional zu Madrid: Studien zur Beatus-Illustration und der spanischen Buchmalerei des 10. Jahrhunderts. Hildesheim: Olms. Klein, Peter K. (1984): Les images de la Genèse de la Bible carolingienne de Bamberg et la tradition des frontispices bibliques de Tours. 77–107. In: Texte et image: Actes du Colloque international de Chantilly (13 au 15 octobre 1982). Paris: Les Belles Lettres. Köhler [Koehler], Wilhelm Reinhold Walter (1930 [1963]): Die karolingischen Miniaturen I: Die Schule von Tours. Tafeln. Berlin: Cassirer. Köhler [Koehler], Wilhelm Reinhold Walter (1933 [1963]): Die karolingischen Miniaturen I: Die Schule von Tours. T.2. Die Bilder. Berlin: Cassirer. Koehler, Wilhelm/Mütherich, Florentine (1994): Die karolingischen Miniaturen VI.1. Die Schule von Reims. Erster Teil. Von den Anfängen bis zur Mitte des 9. Jahrhunderts. Text. Berlin: Deutscher Verlag für Kunstwissenschaft. Koehler, Wilhelm/Mütherich, Florentine (1999): Die karolingischen Miniaturen VI.2. Die Schule von Reims. Zweiter Teil. Von der Mitte bis zum Ende des 9. Jahrhunderts. Text. Berlin: Deutscher Verlag für Kunstwissenschaft. Koshi, Koichi (1973): Die Genesisminiaturen in der Wiener “Histoire universelle” . Vienna: Holzhausen. Lockett, Leslie (2002): An integrated re-examination of the dating of Oxford, Bodleian Library, Junius 11. In: Anglo-Saxon England 31, 141–173. Lucas, Peter J. (1980): MS Junius 11 and Malmesbury. In: Scriptorium 34, 197–220. Lucas, Peter J. (1981): MS Junius 11 and Malmesbury (II). In: Scriptorium 35, 3–22. Lucas, Peter J. (1992): Loyalty and obedience in the Old English Genesis A and the interpolation of Genesis B unto Genesis A. In: Neophilologus 76, 121–135. McKitterick, Rosamond (1990): Text and image in the Carolingian world. In: McKitterick, Rosamond (ed.), The uses of literacy in early mediaeval Europe, 297–318. Cambridge: Cambridge University Press. Menner, Robert J. (1951): The date and dialect of Genesis A 852–2936 (Part III). In: Anglia 70, 285–294. Mittman, Asa Simon/Kim, Susan M. (2015): Locating the devil “her” in MS Junius 11. In: Gesta 54, 3–25. Mütherich, Florentine/Gaehde, Joachim E. (1976): Carolingian painting. New York: Braziller. Neuss, Wilhelm (1931): Die Apokalypse des hl. Johannes in der altspanischen und altchristlichen Bibel-Illustration. I. Text. II. Tafeln. Münster: Aschendorff. Nordenfalk, Carl (1936): Beiträge zur Geschichte der Turonischen Buchmalerei. In: Acta Archaeologica 7, 281–304. Nordenfalk, Carl (1938): Die spätantiken Kanontafeln: Kunstgeschichtliche Studien über die eusebianische Evangelien-Konkordanz in den vier ersten Jahrhunderten ihrer Geschichte. Göteborg: Oscar Isacsons Boktryckeri. Nordenfalk, Carl (1971): Noch eine turonische Bilderbibel. In: Autenrieth, Johanne/Brunhölzl, Franz (eds.), Festschrift Bernhard Bischoff zu seinem 65. Geburtstag: Dargebracht von Freunden, Kollegen und Schülern, 153–163. Stuttgart: Hiersemann.

Secondary sources 

 325

Ohlgren, Thomas H. (1972a): The illustrations of the Caedmonian Genesis: Literary criticism through art. In: Medievalia et Humanistica: Studies in Medieval and Renaissance Culture 3, 199–212. Ohlgren, Thomas H. (1972b): Visual language in the Old English Cædmonian Genesis. In: Visible Language 6, 253–276. Ohlgren, Thomas H. (1975): Some new light on the Old English Caedmonian Genesis. In: Studies in Iconography 1, 38–73. Ohlgren, Thomas H. (1991): Anglo-Saxon art: Texts and contexts. (Old English Newsletter, Subsidia 17) Amherst, MA: Department of English, University of Massachusetts, Amherst. Ohlgren, Thomas H. (1992): Anglo-Saxon textual illustration: Photographs of sixteen manuscripts with descriptions and index. Kalmazoo, MI: Medieval Institute Publications, Western Michigan University. Pächt, Otto (1943): A Giottesque episode in English mediaeval art. In: Journal of the Warburg and Courtauld Institutes 6, 51–70. Pope, John C. (1998): Eduard Sievers (1850–1932). In: Damico, Helen (ed.), Medieval scholarship: Biographical studies on the formation of a discipline. Vol. 2: Literature and philology, 177–199. New York: Garland. Raw, Barbara C. (1953): The story of the fall of man and of the angels in the MS Junius 11 and the relationship of the manuscript illustrations to the text. MA thesis, University of London. Raw, Barbara C. (1955): The drawing of an angel in MS 28, St. John’s College, Oxford. In: Journal of the Warburg and Courtauld Institutes 18, 318–319. Raw, Barbara C. (1976): The probable derivation of most of the illustrations in Junius 11 from an illustrated Old Saxon Genesis. In: Anglo-Saxon England 5, 133–148. Raw, Barbara C. (1984): The construction of Oxford, Bodleian Library, Junius 11. In: Anglo-Saxon England 13, 187–207. Raw, Barbara C. (1997): Trinity and incarnation in Anglo-Saxon art and thought. Cambridge: Cambridge University Press. Reinhard, Ben (2010): The opening image of MS Junius 11. In: Old English Newsletter 42, 15–25; http://www. oenewsletter.org/OEN/print.php/essays/reinhard42_3/; last access September 12, 2022. Rosenthal, Jane (1974): The historiated canon tables of the Arenberg Gospels (New York, Pierpont Morgan Library MS. 869). Ph.D. dissertation, Columbia University. Schapiro, Meyer (1973): Words and pictures: On the literal and the symbolic in the illustration of a text. The Hague: Mouton. [Reproduced in: Schapiro, Meyer, Words, script, and pictures: Semiotics of visual language, 9–114. New York: Braziller, 1996] Schramm, Percy Ernst (1983 [1928]): Die deutschen Kaiser und Könige in Bildern ihrer Zeit 751–1190. New edition. Munich: Prestel. Schroeder, Larry D./Sjoquist, David L./Stephan, Paula E. (2017): Understanding regression analysis: An introductory guide. 2nd edn. Los Angeles: Sage. Sievers, Eduard (1875): Der Heliand und die angelsächsische Genesis. Halle: Lippert. Sievers, Eduard (1924): Ziele und Wege der Schallanalyse: Zwei Vorträge. In: Stand und Aufgaben der Sprachwissenschaft: Festschrift für Wilhelm Streitberg, 65–111. Heidelberg: Winter. Sievers, Eduard (1929): Cædmon und Genesis. In: Britannica: Max Förster zum sechzigsten Geburtstage, 57–84. Leipzig: Tauchnitz. Suzuki, Seiichi (2000): The Quoit Brooch Style and Anglo-Saxon settlement: A casting and recasting of cultural identity symbols. Woodbridge: Boydell. Suzuki, Seiichi (2008): Anglo-Saxon button brooches: Typology, genealogy, chronology. Woodbridge: Boydell. Suzuki, Seiichi (2014a): The meters of Old Norse eddic poetry: Common Germanic inheritance and North Germanic innovation. Berlin: De Gruyter. Suzuki, Seiichi (2014b): Metrical positions and their linguistic realisations in Old Germanic metres: A typological overview. In: Studia Metrica et Poetica 1, 9–38.

326 

 Bibliography

Suzuki, Seiichi (2023): The miniatures and meters of the Old English Genesis, MS Junius 11. Vol. 2: The metrical organization of the Old English Genesis: The Anglo-Saxon foundations and Old Saxon adaptation. Berlin: De Gruyter. Temple, Elżbieta (1976): Anglo-Saxon manuscripts 900–1066. London: Miller. Tronzo, William (1983): The Hildesheim doors: An iconographic source and its implications. In: Zeitschrift für Kunstgeschichte 46, 357–366. Waetzoldt, Stephan (1964): Die Kopien des 17. Jahrhunderts nach Mosaiken und Wandmalereien in Rom. Vienna: Schroll. Weitzmann, Kurt (1947 [1970]): Illustrations in roll and codex: A study of the origin and method of text illustration. Princeton, NJ: Princeton University Press. Weitzmann, Kurt (1955): Observations on the Cotton Genesis fragments. In: Weitzmann, Kurt (ed.), Late classical and mediaeval studies in honor of Albert Mathias Friend, Jr., 112–131. Princeton, NJ: Princeton University Press. Weitzmann, Kurt (1984): The Genesis mosaics of San Marco and the Cotton Genesis miniatures. In: Demus 1984, 105–142. Weitzmann, Kurt/Bernabò, Massimo (1999): The illustrations in the manuscripts of the Septuagint. Vol. 2: Octateuch. The Byzantine Octateuches: text and plates. Department of Art and Archaeology, Princeton University. Williams, John (1987): Tours and the early medieval art of Spain. In: Bjurström, Per/Hökby, Nils-Göran/ Mütherich, Florentine (eds), Florilegium in honorem Carl Nordenfalk octogenarii contextum, 197–208. Stockholm: Nationalmuseum. Williams, John (1994–2003): The illustrated Beatus: A corpus of the illustrations of the Commentary on the Apocalypse. 5 vols. Vol. 1: Introduction. Vol. 2: The 9th and 10th centuries. Vol. 3: The 10th and 11th centuries. Vol. 4: The 11th and 12th centuries. Vol. 5: The 12th and 13th centuries. London: Miller.

Index Index of authors Alexander, J. J. G. 21 Anlezark, Daniel 30, 34, 35, 43, 64, 97, 105, 157, 158, 166, 203, 213, 240, 245, 246 Austin, J. L. 124 Behaghel, Otto/Taeger, Burkhard 2 Bergman, Robert P. 15, 102, 106, 111 Blum, Pamela Z. 93, 94, 111, 117, 120–125 Broderick, Herbert R. VIII, 4, 18, 21, 23, 24, 27, 28, 29, 30, 31, 34, 38, 46, 51, 52, 59, 63, 64, 66, 68, 76, 78, 84, 86, 93, 94, 97, 98, 99, 100, 102, 110, 111, 113, 114, 121, 123, 135, 140, 142, 144, 154, 159, 162, 163, 167, 177, 185, 186, 187, 188, 192, 194, 200, 201, 209, 215, 219, 224, 226, 230, 237, 239, 243, 244, 245, 246, 247, 248, 279, 286, 289, 311 Büchsel, Martin/Kessler, Herbert L./Müller, Rebecca 15, 99, 103, 109 Cardinali, Marco 16, 40 Coogan, Michael D. (New Revised Standard Version) 27, 28, 32, 78, 80, 86, 88, 135 Cook, Robin 16 Dell’Acqua, Francesca et al. 15, 103 Demus, Otto 15, 99, 103, 104, 237 DeWald, Ernest T. 27, 76, 84, 86 Diebold, William J. 40, 111, 265 Doane, A. N. 1, 2, 3, 5, 6, 11, 12, 136, 152, 157, 163, 298, 299, 300, 301, 311 Dorfman, Eugene 30 Dufrenne, Suzy 27, 80 Dutton, Paul Edward/Kessler, Herbert L. 40, 71, 265 Ericksen, Janet Schrunk 159 Feinstein, Charles H./Thomas, Mark 304 Finnegan, Robert 94 Fischer, Bonifatius 16 Fulk, R. D./Bjork, Robert E./Niles, John D. 1, 6 Gaehde, Joachim E. 16, 211 Gameson, Richard 159 Ganz, Peter 11 Garipzanov, Ildar H. 40, 111 Gneuss, Helmut/Lapidge, Michael 1 Gollancz, Sir Israel 1, 28, 29, 30, 31, 35, 59, 60, 64, 66, 67, 68, 93, 94, 99, 100, 101, 116, 127, 157, 158, 159, 165, 185, 186, 188, 223, 237 Green, Rosalie B. 15, 210, 226 https://doi.org/10.1515/9783110788068-015

Green, Rosalie/Evans, Michael/Bischoff, Christine/ Curschmann, Michael 15, 147, 209, 211, 297 Grice, Herbert Paul 123 Hamano, Akihiro 15 Henderson, George 4, 97–98, 121, 245, 248 Hilmo, Maidie 135, 224, 261 Horst, Koert van der 80 Hubert, Jean/Porcher, Jean/Volbach, W. F. 23, 167 Jolly, Penny Howell 175, 178, 179, 227, 228 Joslin, Mary Coker/Watson, Carolyn Coker Joslin 120 Karkov, Catherine E. VIII, IX, 1, 2, 27, 32, 34, 64, 66, 132, 135, 158, 161, 162, 166, 185, 237, 239 Kaspersen, Søren 16, 140, 198, 265 Ker, N. R. 1 Kessler, Herbert L. VIII, 15, 16, 22, 28, 40, 45, 70, 73, 84, 90, 102, 109, 131, 133, 136, 140, 148, 167, 174, 183, 195, 198, 199, 209, 210, 211, 212, 228, 234, 241, 242, 264, 279, 294 Keynes, Simon/Lapidge, Michael 301 Klaeber, Fr. 9 Klein, Peter K. 16, 241, 242, 279, 287 Köhler [Koehler], Wilhelm Reinhold Walter 15, 16, 22, 23, 40, 41, 45, 49, 109 Koehler, Wilhelm Reinhold Walter/Mütherich, Florentine 16, 27 Koshi, Koichi 276, 277 Krapp, George Philip 1 Lockett, Leslie 1 Lucas, Peter J. 1, 18, 100, 122, 311 McKitterick, Rosamond 80 Menner, Robert J. 11 Mittman, Asa Simon/Kim, Susan M. 59, 64 Mütherich, Florentine/Gaehde, Joachim E. 41, 109, 114 Neuss, Wilhelm 287 Nordenfalk, Carl 16, 118 Ohlgren, Thomas H. 28, 29, 30–31, 34, 43, 52, 59, 63, 64, 68, 94, 157, 159, 185, 186, 188, 193, 217, 223, 237, 286, 288, 289, 299 Pächt, Otto 120 Pope, John C. 11 Raw, Barbara C. VIII, IX, 1, 17–19, 27, 28, 29, 30, 34, 51, 52, 57, 59, 63, 66, 86, 93, 95–97, 102, 106, 107, 114, 142, 158, 159, 163, 167, 169, 170, 180,

328 

 Index

181, 183, 188, 192, 194, 215, 217, 246, 281, 298, 300 Reinhard, Ben 27, 51 Rosenthal, Jane 64, 65, 84 Schäfer, Frank 15 Schapiro, Meyer 37, 255 Schramm, Percy Ernst 40 Schroeder, Larry D./Sjoquist, David L./Stephan, Paula E. 304 Sievers, Eduard 2, 11, 301, 310

Suzuki, Seiichi VIII, IX, 1, 8, 9, 10, 116 Temple, Elżbieta 1, 21, 48, 293, 294 Tronzo, William 140 Waetzoldt, Stephan 15 Weitzmann, Kurt 102, 174, 178, 276, 313 Weitzmann, Kurt/Bernabò, Massimo 121, 122 Weitzmann, Kurt/Kessler, Herbert L. VIII, 14, 15, 171, 174, 178, 180, 253, 276, 277, 279 Williams, John 52, 286, 287, 288, 289, 291 Williams, John/Shailor, Barbara A. 286

Index of subjects Adam and Eve cycle 2, 7, 15, 16, 18, 199, 254, 255, 258–261, 267, 269, 273–274, 278–280, 301, 313 Æthelwulf, King 300 Alfred, King 300–301 alliteration 8–11, 301–311 analogical extension 50, 150, 199, 314 analogy 142, 154, 221, 237, 238, 259, 266 Anglian Genesis 2, 301, 308, 311, 312, 313, 317. See also Genesis A; Old English Genesis Antwerp, Plantin-Moretus Museum, M 17.4 (Carmen Paschale) 21 Asser’s Life of King Alfred 301 back-formation 259 Bamberg, Staattsbibliothek, Msc. Bibl. 1 (Bamberg Bible) 16, 131, 137–140, 147–153, 167, 192–198, 208–215, 231–235, 241–243, 246, 248, 252–253, 263, 267–274, 278–280, 284–285, 288, 315, 317, Pl. 23 Bamberg, Staatsbibliothek, Msc. Bibl. 22 (Bamberg Daniel Commentary) 27 Bamberg, Staatsbibliothek, Msc. Bibl. 76 (Bamberg Commentaries/Isaias Glossatus) 57 Beatus commentary 52–53, 286–292 Berlin, Staatliche Museen zu Berlin, ivory plaque (Ident. Nr.: 589) 102 Berlin, Staatsbibliothek zu Berlin, Ms. theol. lat. fol. 733 (Prüm Gospels) 109 Budapest, Museum of Fine Arts, Salerno ivories, Creation of Birds and Fish Pl. 54. See also Salerno Burgo de Osma, Archivo de la Catedral, Cod. 1 (Osma Beatus) 287, 290. See also Beatus commentary Canon table 106, 110, 116, 117–119, 284 Charles the Bald 16, 19, 40–42, 47, 114, 153, 264, 266, 300 Christ and Satan 1, 2 Cotton Genesis (family) 14–16, 102, 112, 121–122, 131, 136, 148, 151, 160, 170–180, 195, 197, 200, 211, 212, 215, 217, 226, 230, 233–234, 237, 241–248, 251, 253, 267, 269, 271, 276–278, 279, 285, 299, 314, 315, 316 Creation cycle 2, 7, 93–101, 102–125, 174, 261–262, 313, 314

Daniel 1, 2, 6 (distinctive or oppositional) feature (or parameter) 23–27, 35–36, 37, 38–40, 80–81, 148–151, 201–208, 241–249, 251, 255–266, 314–315 Épernay, Bibliothèque Municipale, Ms. 1 (Ebbo Gospels) 114, 121 Escorial, Biblioteca del Monasterio, &. II.5 (Escorial Beatus) 287, 290. See also Beatus commentary Evangelist at work 109, 114, 253 Exodus, Book of 255 Exodus 1, 2, 6 Ezekiel, Book of 27 Fall of the Rebel Angels cycle 1, 2, 7, 18, 52, 59–91, 135, 252, 254, 256–258, 266–267, 269, 274, 275, 290, 301, 313 frieze 116–119, 134, 148–152, 215, 283–284 Genesis, Book of 1, 3, 52, 93, 96, 99, 135, 151, 163, 165, 167, 233, 243–244, 248, 267, 269, 279, 296 Genesis A (1–234, 852–2936) 1–14, 298–312, 313–317. See also Anglian Genesis; Genesis B; Old English Genesis; Old English/Old Saxon Genesis Genesis A1 (1–234, 852–966) 9–13, 19, 301–312, 313–317. See also Anglian Genesis; Genesis A; Genesis A2; Old English Genesis; Old English/ Old Saxon Genesis Genesis A2 (967–2936) 9–13, 301–312, 313–317. See also Anglian Genesis; Genesis A; Genesis A1; Old English Genesis; Old English/Old Saxon Genesis Genesis B (235–851) 1–14, 19, 298–312, 313–317. See also Genesis A1; Old English Genesis; Old English/Old Saxon Genesis; Old Saxon Genesis Girona, Museu de la Catedral de Girona, Num Inv. 7(11) (Girona Beatus) 287, 288. See also Beatus commentary God/Christ in Majesty 21, 23, 28, 45, 52, 109, 153–155, 264–265, 270, 281, 288, 292–295, 317 Grimbald of St. Bertin’s 301 Hildesheim, Bronze Door of St. Michael’s Church 16, 140, 175, 198, 209, 211, 217, 233–235 Isaiah, Book of 27, 28, 32, 51, 52, 57

330 

 Index

John the Old Saxon 300 (pre-)Junius artist 16, 20, 82, 99, 111, 113, 116, 119, 122, 124, 130, 131, 146, 150–151, 167, 169, 195, 214, 230, 231, 235, 246–248, 254, 264, 265, 267–276, 282–285, 313, 316 Kings, Book of 27 King on the throne 109, 110, 114, 116 Klagenfurt, Kärntner Landesarchiv, Cod. GV 6/19 (Millstatt Genesis) 15, 131, 133, 136, 147–153, 170, 171, 177, 195, 210–212, 233–234, 238, 241–246, 248–249, 252, 260, 270, 284, 294, 299, Pls. 58–67 linear regression 302–310 London, British Library, Add MS 10546 (Grandval Bible) 16, 48, 109, 131, 133, 136, 141, 147, 153, 159, 167, 169–182, 192–198, 209–211, 217, 233–235, 241–243, 253, 260, 263, 267–276, 278, 279, 288, Pls. 24–25 London, British Library, Add. MS. 11695 (Silos Beatus) 287. See also Beatus commentary London, British Library, Cotton MS Claudius B IV (Ælfric’s Old English Hexateuch) 2 London, British Library, Cotton MS Otho B VI (Cotton Genesis) 14 London, British Library, Egerton MS 1894 (Egerton Genesis) 120–122 London, British Library, Harley MS 2788 (Harley Golden Gospels) 117–118 Madrid, Biblioteca Nacional, MS Vitrina 14-2 (Facundus Beatus) 287, Pl. 83. See also Beatus commentary Madrid, Real Academia de la Historia, Cod. 33 (San Millán Beatus) 287, 290. See also Beatus commentary Manchester, John Rylands University Library, MS lat. 8 (Rylands Beatus) 287. See also Beatus commentary metrical organization 4–11, 108, 301–310, 313 Munich, Bayerische Staatsbibliothek, Clm 14000 (Codex Aureus of St. Emmerams) 114, 116, 121 Munich, Bayerische Staatsbibliothek, Clm 14345 (Epistulae Pauli) 84 Nancy, Trésor de la cathédrale, Cathédrale de Notre-Dame de l’Annonciation, St. Gauzelin Gospels 23

narrative progression, left to right vs. right to left 69, 71, 80, 86, 105, 107, 128, 131, 139, 140, 158–160, 162, 164, 170, 176, 178, 185, 186, 187, 188, 221, 240, 273–275, 316 narrative progression, top to bottom vs. bottom to top 69, 101, 105–108, 141, 150, 217–218 New York, Metropolitan Museum of Art, Salerno ivories, Plaque with God Creating the Animals Pl. 55. See also Salerno New York, Pierpont Morgan Library, MS M.429 (Huelgas Beatus) 287, 290. See also Beatus commentary New York, Pierpont Morgan Library, MS M.644 (Morgan Beatus) 52, 53, 286, 288, 289, 291. See also Beatus commentary Octateuch 120–125, 215 Old English Genesis (1–2936) 1–14, 15, 298–312, 313–317. See also Genesis A; Genesis B Old English/Old Saxon Genesis (1–966) 11, 12, 19, 163, 167, 178, 237, 255, 266, 269, 280, 291, 307–312, 313–317. See also Genesis A; Genesis A1; Genesis A2; Genesis B; Old Saxon Genesis Old Saxon Genesis (Vatican or Touronian) 2, 17–19, 160, 280, 290, 291, 292, 297, 298–312, 313–317. See also Genesis B; Old English/Old Saxon Genesis Oxford, Bodleian Library, MS Junius 11 1–14, Pls. 1–22, 71. See also Pii through P46 Pii Pl. 1; 4, 7, 8, 9, 12, 17, 18, 19, 20, 21–57, 105–106, 141, 143, 144, 147, 153, 154, 225–226, 239, 252, 254–255, 256, 257, 258, 261–262, 263, 265, 266–267, 281–282, 292–295, 314, 317 P2 Pls. 2a, 2b; 4, 7, 12, 14, 17, 18, 21–57, 61, 62, 65–66, 67–68, 86, 101, 105–106, 141, 143–144, 154, 225–226, 239, 247, 252, 254–255, 256, 257, 258, 261–262, 263, 265, 266–267, 274, 275, 281, 286–291, 294, 314, 316–317 P3 Pl. 3; 4, 7, 12, 14, 17, 18, 29, 30, 31, 32, 35–37, 41, 43, 53, 59–91, 101, 106, 119, 130, 131, 134, 135–136, 150, 159, 176, 187, 190, 193, 223, 238, 247, 249, 252, 254, 256, 257, 258, 266–267, 270, 272, 274, 275, 281–285, 286, 314 P6 Pl. 4; 4, 7, 8, 12, 14, 29, 93–125, 142, 145, 146, 152, 201, 202, 249, 253, 261–262, 266, 314 P7 Pl. 5; 4, 8, 12, 14, 29, 93–125, 142, 144–146, 152, 201, 202, 218, 249, 253, 261–262, 266, 281, 283–284, 285, 315

Index of subjects 

P9 Pl. 6; 4, 7, 12, 14, 17, 59, 64–65, 127–140, 159, 161, 165, 170, 176, 199, 218, 221, 239, 246, 253, 261, 266, 268, 269, 272, 273, 274 P10 Pl. 7; 4, 5, 7, 12, 14, 29, 132, 141–155, 159, 163, 164, 165–166, 169, 177–178, 199, 205–208, 221, 223–226, 237, 240, 249, 252, 253, 254, 258, 259, 260, 261, 262, 266, 274 P11 Pl. 8; 4, 7, 8, 12, 14, 115, 119, 132, 141–155, 163, 164, 165–166, 169, 205–208, 223–226, 237, 252, 254, 258, 259, 260, 261, 262, 263, 270, 279, 281, 283–285 P13 Pl. 9; 4, 7, 12, 14, 115, 132, 141–155, 163, 164, 165–166, 169, 186, 188, 205–208, 223–224, 239, 240–241, 252, 254, 259, 260, 261, 263, 279, 281, 283–285 P16 Pl. 10; 4, 7, 12, 14, 35–37, 67, 68, 88, 89–91, 130, 201, 238, 247, 249, 252, 254, 256, 258, 274, 282 P17 Pl. 11; 4, 7, 12, 14, 21, 35–37, 42, 66, 68, 89–91, 129–130, 133, 146–147, 157, 165, 201, 238, 252, 254, 256, 258, 265, 282 P20 Pl. 12; 4, 12, 14, 59, 127–129, 131, 134–135, 157–183, 186–187, 190, 197, 252, 253, 254, 257, 259, 260, 261, 263, 273, 278, 295–296 P24 Pl. 13; 4, 7, 12, 14, 157–183, 186, 187, 189–190, 253, 261, 267, 270, 295–296 P28 Pl. 14; 4, 7, 12, 14, 61, 163, 181, 185–200, 214, 221–222, 223, 235, 238, 253, 261, 270, 273, 296 P31 Pl. 15; 4, 12, 14, 177, 181, 185–200, 201–203, 205, 208–209, 212–215, 217, 218–220, 223, 231–233, 240, 252, 253, 259–261, 263, 267–268, 269–270, 271, 279, 296, 315 P34 Pl. 16; 4, 7, 8, 12, 14, 191, 201–215, 217, 218, 220, 232–233, 252, 253, 254, 259–261, 263, 266, 271 P36 Pl. 17; 4, 7, 8, 12, 14, 59, 165, 191, 201–215, 217, 218, 220, 232, 252, 254, 259, 260, 261, 263 P39 Pl. 18; 4, 7, 12, 14, 191, 201–215, 217, 218, 219–220, 232, 252, 254, 259, 260, 261, 263 P41 Pl. 19; 4, 12, 14, 101, 131, 191, 200, 202, 205–206, 207–208, 213, 217–235, 238, 252, 253, 254, 259–260, 261, 263, 267–268, 270, 272, 273–274, 276–278, 279, 292, 316 P44 Pl. 20; 4, 12, 14, 132, 185, 200, 202, 205–208, 217–235, 237, 247, 249, 252, 253, 261, 267, 268, 272, 273–274, 276–278, 292, 316 P45 Pl. 21; 4, 7, 12, 14, 177, 186, 205–208, 217–235, 237–249, 252, 254, 260, 261, 263, 274, 297 P46 Pl. 22; 4, 7, 8, 9, 12, 14, 17, 19, 133, 186, 187–188, 218, 237–249, 253, 261, 266, 279, 295, 297, 317

 331

Paris, Bibliothèque de L’Arsenal Ms. 1171 (St. Aure Gospel) 153, 154 Paris, Bibliothèque Nationale, Ms. lat. 1 (Vivian Bible) 16, 18, 19, 22, 23, 28, 39–53, 54, 55, 56, 61, 69–91, 109, 111, 118, 121, 130–140, 141, 153–155, 159, 167, 178, 192–198, 209–211, 214, 217, 233–234, 241–243, 252–253, 258, 264–265, 267–275, 278, 281, 282, 288, 290, 292–295, 298, 300, 315–317, Pls. 26–32. See also Vivian 423r Paris, Bibliothèque Nationale, Ms. lat. 6 (Roda Bible) 18, 51–52, 279 Paris, Bibliothèque Nationale, Ms. lat. 261 (Le Mans Gospels) 22, 48, 109 Paris, Bibliothèque Nationale, Ms. lat. 266 (Lothair Gospels) 22, 41, 48, 109 Paris, Bibliothèque Nationale, Ms. lat. 1152 (Psalter of Charles the Bald) 47 Paris, Bibliothèque Nationale, Ms. lat. 6401 (Apotheosis of Boethius) 21–23, 42, 48–49, 293–294, Pl. 73 Paris, Bibliothèque Nationale, Ms. lat. 8850 (Gospels of St. Médard de Soissons) 109 Paris, Bibliothèque Nationale, Ms. lat. 8878 (Saint-Sever Beatus) 287, Pl. 84. See also Beatus commentary Paris, Bibliothèque Nationale, Ms. lat. 9385 (Dufay Gospels) 22, 48, 109 Paris, Musée du Louvre, Leaf of the Diptych of the Consul Areobindus 167 pictorial organization 18, 74, 147, 177, 223–224, 232, 237–238, 247, 249, 251–266, 271–275, 298–299, 314–317 picture, placement of, in MS Junius 11 4–8, 12–14 Psalms, Book of 27, 78, 80, 86, 88 Revelation, Book of 28, 52, 53, 286, 288, 290 Rome, San Paolo fuori le Mura (San Paolo Bible) 15, 16, 19, 22, 23, 28, 40, 42, 45, 48, 51, 69–91, 109, 131, 136, 140, 141, 153–154, 159, 183, 197–199, 209–212, 217, 233–234, 241–243, 252–253, 258, 264–274, 278, 282, 298, Pls. 33–37 Rome, San Paolo fuori le Mura frescoes (destroyed). See Vatican, Biblioteca Apostolica, Cod. Barb. lat. 4406 Salerno, Museo del Duomo, Salerno ivories, 15, 99, 102–119, 122, 152, 171, 175–176, 241–243, 253, Pls. 52–57. See also Budapest; New York

332 

 Index

Samuel, Book of 27 section, a metrical unit 2–3, 6, 12–14, 301–311 Strasbourg, Bibliothèque de la Ville, Hortus Deliciarum (destroyed) 15, 147, 171–173, 177, 204, 209–212, 233–234, 238, 241–244, 248, 263, 268, 271, 297, Pls. 68–70 Touronian Bibles 14–16, 17–19, 84, 153–154, 199, 241–248, 251–253, 264–265, 267–276, 278–280, 298–300, 314–317 Turin, Biblioteca Nazionale Universitaria, Sgn. I.II.1 (Turin Beatus) 287, 290. See also Beatus commentary Urgell, Museu Diocesá de La Seu d’Urgell, Num Inv. 501 (Urgell Beatus) 287. See also Beatus commentary Utrecht, Universiteitsbibliotheek, MS Bibl. Rhenotraiectinae I Nr 32 (Utrecht Psalter) 17–19, 27, 51, 76–82, 84–91, 110, 118–119, 147–152, 281–285, 298, 315, Pls. 74–82 Valladolid, Biblioteca de la Universidad, MS 433 (Valladolid Beatus) 52, 286. See also Beatus commentary Vatican, Biblioteca Apostolica, Cod. Barb. lat. 4406 (San Paolo fuori le Mura frescoes) 15, 233–234, 241–243

Vatican, Biblioteca Apostolica, Pal. lat. 1447 (containing Old Saxon Genesis fragments) 2. See also Old Saxon Genesis Vatican, Biblioteca Apostolica, Vat. gr. 747 (Octateuch) 121 Venice, Basilica di San Marco. Interior. Mosaics. Atrium. West Arm. South Cupola (The Creation Cupola) 15, 52, 99, 102–117, 122, 131–134, 136, 138–139, 147–153, 167–171, 174–181, 183, 192–198, 200, 202, 205, 209–212, 217, 219, 226–234, 237, 238, 241–245, 253, 260, 267–268, 270–271, 276–296, 297, 298, 315–316, Pls. 38–51, 72 Vercelli, Bibliotheca Capitolare (Rotulus) 84 Vienna, Österreichische Nationalbibliothek, Cod. 652 (Hrabanus Maurus’ De laudibus sanctae crucis) 84 Vienna, Österreichische Nationalbibliothek, Cod. 2576 (Histoire Universelle en Prose) 276–278 Vienna, Österreichische Nationalbibliothek, Cod. Theol. gr. 7 276 Vivian 423r Pl. 32; 18, 22, 28, 39, 40–57, 61, 75, 109, 111, 252, 264–265, 275, 290, 292–294, 300, 315, 316, 317 Zurich, Zentralbibliothek, Ms Car C1 (Bible with prologs and capitula) 288

Index of verses inscription (p. 2): 23, 257 inscriptions (p. 3): 59 inscription (p. 6): 93, 98, 120, 122–124 inscription (p. 7): 94, 97, 122–124 inscriptions (p. 9): 127 Genesis A 1a: 5 1a–8a: 34 1a–12a: 34 1a–77b: 4 12b–21b: 34 18b–21b: 30 22a–28a: 34 31b–36a: 34 36b–38b: 43 49a: 5 103a–104a: 97 103a–106a: 43 103a–110a: 105 114a: 109 114b: 109 114b–135b: 108 124b–125a: 114 135a: 5 135b: 108 143b–153b: 96 147b–150a: 97 185b: 132, 133, 166 186a: 5 205b: 5 206a: 5 210a: 5 234b: 3 852a: 3 852a–854b: 191 852a–917b: 101 872a: 5

903a–938b: 218, 221 939a: 5 945a–947b: 245, 248 952a: 5 955a–957b: 240 955a–961b: 246 964b: 9 966b: 9 967a: 9 972a: 5 994a: 5 1036a: 5 1063a: 5 1101a: 5 1143a: 5 1159a: 5 1167a: 5 1179a: 5 1215a: 5 1225a: 5 1237a: 5 1314a: 5 1327a: 5 1364a: 5 1497a: 5 1584b–1588a: 13 1588b: 5 1602a: 5 1697a: 5 1719a: 5 1767a: 5 1830a: 5 Genesis B 235a: 3 235a–236b: 3 246a: 5 246a–258b: 34–35 246a–337b: 4 252b–256a: 29 259a–264b: 34 259a–271b: 34 261b–266a: 31

271b–272a: 31 274a: 61 283b–289a: 64 292a–295a: 31, 34 309b: 5 313a–441a: 157 325a: 5 338a–440a: 134 365a–368a: 127 389a: 5 395a–397a: 127 441a: 163 442a: 163 442a–453b: 157 447a: 134 454a–459b: 135 454a–460b: 157 454a–490b: 157, 182 454a–491b: 169 460a–460b: 164 467a–485b: 164 491a: 5 491a–491b: 164 491a–494b: 135, 158 491a–546b: 182 495a–546b: 161, 166 495a–599b: 193 497b–499a: 165, 166 538a–539b: 166 599a: 5 663b: 5 717b–719b: 187 732b: 5 765a: 5 765b–783a: 213 777b–781a: 213 783b–784a: 213 841b: 3 842a: 5 845a–845b: 202, 213 851b: 3

Plates

Plate 1. Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. ii, God Enthroned. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0). https://doi.org/10.1515/9783110788068-016

338 

 Plates

Plate 2a. Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 2, God Enthroned, Facing Lucifer. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).

Plates 

Plate 2b. Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 2, God Enthroned, Facing Lucifer (detail). Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).

 339

340 

 Plates

Plate 3. Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 3, Fall of the Rebel Angels. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).

Plates 

 341

Plate 4. Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 6, Creation. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).

342 

 Plates

Plate 5. Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 7, Creation. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).

Plates 

 343

Plate 6. Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 9, Creation of Eve. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).

344 

 Plates

Plate 7. Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 10, God Blesses Adam and Eve. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).

Plates 

 345

Plate 8. Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 11, God Beholds His Creations. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).

346 

 Plates

Plate 9. Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 13, Adam and Eve in Paradise. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).

Plates 

 347

Plate 10. Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 16, Fall of the Rebel Angels. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).

348 

 Plates

Plate 11. Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 17, Fall of the Rebel Angels. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).

Plates 

 349

Plate 12. Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 20, The Devil Leaves Hell/Temptation of Adam. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).

350 

 Plates

Plate 13. Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 24, Temptation of Adam. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).

Plates 

 351

Plate 14. Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 28, Adam Withstands the Temptation of the Devil/Eve Succumbs. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).

352 

 Plates

Plate 15. Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 31, Eve Tempts Adam/Adam and Eve Pray in Remorse. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).

Plates 

 353

Plate 16. Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 34, Adam and Eve Cover Themselves. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).

354 

 Plates

Plate 17. Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 36, Adam and Eve Cover Themselves/The Devil Returns to Hell. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).

Plates 

 355

Plate 18. Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 39, Adam and Eve Cover Themselves. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).

356 

 Plates

Plate 19. Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 41, God Calls to Adam and Eve/Judgment of the Serpent. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).

Plates 

 357

Plate 20. Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 44, Judgment of Adam and Eve. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).

358 

 Plates

Plate 21. Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 45, Expulsion. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).

Plates 

 359

Plate 22. Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 46, Expulsion. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).

360 

 Plates

Plate 23. Bamberg, Staatsbibliothek, Msc. Bibl. 1 (Bamberg Bible), fol. 7v, Genesis frontispiece (licensed under CC BY-SA 4.0).

Plates 

Plate 24. London, British Library, Add MS 10546 (Grandval Bible), fol. 5v, Genesis frontispiece. © The British Library Board.

 361

362 

 Plates

Plate 25. London, British Library, Add MS 10546 (Grandval Bible), fol. 352v, God in Majesty. © The British Library Board.

Plates 

 363

Plate 26. Paris, Bibliothèque Nationale, MS lat. 1 (Vivian Bible), fol. 3v, St. Jerome frontispiece. Source gallica.bnf.fr/Bibliothèque Nationale de France (licensed under CC BY 4.0).

364 

 Plates

Plate 27. Paris, Bibliothèque Nationale, MS lat. 1 (Vivian Bible), fol. 10v, Genesis frontispiece. Source gallica. bnf.fr/Bibliothèque Nationale de France (licensed under CC BY 4.0).

Plates 

 365

Plate 28. Paris, Bibliothèque Nationale, MS lat. 1 (Vivian Bible), fol. 130v (detail), God in Majesty. Source gallica.bnf.fr/Bibliothèque Nationale de France (licensed under CC BY 4.0).

366 

 Plates

Plate 29. Paris, Bibliothèque Nationale, MS lat. 1 (Vivian Bible), fol. 326r, Canon Table. Source gallica.bnf.fr/ Bibliothèque Nationale de France (licensed under CC BY 4.0).

Plates 

 367

Plate 30. Paris, Bibliothèque Nationale, MS lat. 1 (Vivian Bible), fol. 329v, God in Majesty. Source gallica.bnf. fr/Bibliothèque Nationale de France (licensed under CC BY 4.0).

368 

 Plates

Plate 31. Paris, Bibliothèque Nationale, MS lat. 1 (Vivian Bible), fol. 386v, Epistles frontispiece. Source gallica.bnf.fr/Bibliothèque Nationale de France (licensed under CC BY 4.0).

Plates 

 369

Plate 32. Paris, Bibliothèque Nationale, MS lat. 1 (Vivian Bible), fol. 423r, Dedication frontispiece. Source gallica.bnf.fr/Bibliothèque Nationale de France (licensed under CC BY 4.0).

370 

 Plates

Plate 33. Rome, San Paolo fuori le Mura, Bibbia carolingia (San Paolo Bible), fol. 3v, St. Jerome frontispiece. Reproduction: Cardinali (2009: 23), © Edizioni Abbazia San Paolo.

Plates 

 371

Plate 34. Rome, San Paolo fuori le Mura, Bibbia carolingia (San Paolo Bible), fol. 8v, Genesis frontispiece. Reproduction: Cardinali (2009: 25), © Edizioni Abbazia San Paolo.

372 

 Plates

Plate 35. Rome, San Paolo fuori le Mura, Bibbia carolingia (San Paolo Bible), fol. 117r (detail), Prophets frontispiece. Reproduction: Cardinali (2009: 43), © Edizioni Abbazia San Paolo.

Plates 

 373

Plate 36. Rome, San Paolo fuori le Mura, Bibbia carolingia (San Paolo Bible), fol. 259v, God in Majesty. Reproduction: Cardinali (2009: 53), © Edizioni Abbazia San Paolo.

374 

 Plates

Plate 37. Rome, San Paolo fuori le Mura, Bibbia carolingia (San Paolo Bible), fol. 310v, Epistles frontispiece. Reproduction: Cardinali (2009: 67), © Edizioni Abbazia San Paolo.

Plates 

 375

Plate 38. Venice, Basilica di San Marco. Interior. Mosaics. Atrium. West Arm. South Cupola (The Creation Cupola), God Forming Adam, SM0767. Ekkehard Ritter, North Adriatic Project Fieldwork Records and Papers, 1975–1979, MS.BZ.009, Dumbarton Oaks, Trustees for Harvard University, Washington, D.C.

376 

 Plates

Plate 39. Venice, Basilica di San Marco. Interior. Mosaics. Atrium. West Arm. South Cupola (The Creation Cupola), Blessing of the Seventh Day, SM0770. Ekkehard Ritter, North Adriatic Project Fieldwork Records and Papers, 1975–1979, MS.BZ.009, Dumbarton Oaks, Trustees for Harvard University, Washington, D.C.

Plates 

 377

Plate 40. Venice, Basilica di San Marco. Interior. Mosaics. Atrium. West Arm. South Cupola (The Creation Cupola), Animation of Adam, SM0772. Ekkehard Ritter, North Adriatic Project Fieldwork Records and Papers, 1975–1979, MS.BZ.009, Dumbarton Oaks, Trustees for Harvard University, Washington, D.C.

378 

 Plates

Plate 41. Venice, Basilica di San Marco. Interior. Mosaics. Atrium. West Arm. South Cupola (The Creation Cupola), God Introducing Adam into Paradise, SM0775. Ekkehard Ritter, North Adriatic Project Fieldwork Records and Papers, 1975–1979, MS.BZ.009, Dumbarton Oaks, Trustees for Harvard University, Washington, D.C.

Plates 

 379

Plate 42. Venice, Basilica di San Marco. Interior. Mosaics. Atrium. West Arm. South Cupola (The Creation Cupola), Adam Naming the Animals, SM0781. Ekkehard Ritter, North Adriatic Project Fieldwork Records and Papers, 1975–1979, MS.BZ.009, Dumbarton Oaks, Trustees for Harvard University, Washington, D.C.

380 

 Plates

Plate 43. Venice, Basilica di San Marco. Interior. Mosaics. Atrium. West Arm. South Cupola (The Creation Cupola), God Creating Eve, SM0784. Ekkehard Ritter, North Adriatic Project Fieldwork Records and Papers, 1975–1979, MS.BZ.009, Dumbarton Oaks, Trustees for Harvard University, Washington, D.C.

Plates 

 381

Plate 44. Venice, Basilica di San Marco. Interior. Mosaics. Atrium. West Arm. South Cupola (The Creation Cupola), God Introducing Adam and Eve, SM0787. Ekkehard Ritter, North Adriatic Project Fieldwork Records and Papers, 1975–1979, MS.BZ.009, Dumbarton Oaks, Trustees for Harvard University, Washington, D.C.

382 

 Plates

Plate 45. Venice, Basilica di San Marco. Interior. Mosaics. Atrium. West Arm. South Cupola (The Creation Cupola), Temptation of Eve, SM0789. Ekkehard Ritter, North Adriatic Project Fieldwork Records and Papers, 1975–1979, MS.BZ.009, Dumbarton Oaks, Trustees for Harvard University, Washington, D.C.

Plates 

 383

Plate 46. Venice, Basilica di San Marco. Interior. Mosaics. Atrium. West Arm. South Cupola (The Creation Cupola), Eve Plucking the Fruit and Giving It to Adam, SM0790. Ekkehard Ritter, North Adriatic Project Fieldwork Records and Papers, 1975–1979, MS.BZ.009, Dumbarton Oaks, Trustees for Harvard University, Washington, D.C.

384 

 Plates

Plate 47. Venice, Basilica di San Marco. Interior. Mosaics. Atrium. West Arm. South Cupola (The Creation Cupola), Adam and Eve Covering Themselves with Fig Leaves, SM0791. Ekkehard Ritter, North Adriatic Project Fieldwork Records and Papers, 1975–1979, MS.BZ.009, Dumbarton Oaks, Trustees for Harvard University, Washington, D.C.

Plates 

 385

Plate 48. Venice, Basilica di San Marco. Interior. Mosaics. Atrium. West Arm. South Cupola (The Creation Cupola), Adam and Eve Hiding from the Presence of God, SM0792. Ekkehard Ritter, North Adriatic Project Fieldwork Records and Papers, 1975–1979, MS.BZ.009, Dumbarton Oaks, Trustees for Harvard University, Washington, D.C.

386 

 Plates

Plate 49. Venice, Basilica di San Marco. Interior. Mosaics. Atrium. West Arm. South Cupola (The Creation Cupola), Adam and Eve Denying Their Guilt, SM0794. Ekkehard Ritter, North Adriatic Project Fieldwork Records and Papers, 1975–1979, MS.BZ.009, Dumbarton Oaks, Trustees for Harvard University, Washington, D.C.

Plates 

 387

Plate 50. Venice, Basilica di San Marco. Interior. Mosaics. Atrium. West Arm. South Cupola (The Creation Cupola), Punishment of Adam and Eve, and the Curse of the Serpent, SM0796. Ekkehard Ritter, North Adriatic Project Fieldwork Records and Papers, 1975–1979, MS.BZ.009, Dumbarton Oaks, Trustees for Harvard University, Washington, D.C.

388 

 Plates

Plate 51. Venice, Basilica di San Marco. Interior. Mosaics. Atrium. West Arm. South Cupola (The Creation Cupola), God Expelling Adam and Eve from Paradise/Adam and Eve’s Labor, SM0800. Ekkehard Ritter, North Adriatic Project Fieldwork Records and Papers, 1975–1979, MS.BZ.009, Dumbarton Oaks, Trustees for Harvard University, Washington, D.C.

Plates 

 389

Plate 52. Salerno, Museo del Duomo, Salerno ivories, The Spirit over the Waters and the Separation of Light and Darkness/Creation of the Firmament. Photo: Kunsthistorisches Institut in Florenz, Max-Planck-­ Institut fle0009789x_p; Photographer: Roberto Sigismondi.

390 

 Plates

Plate 53. Salerno, Museo del Duomo, Salerno ivories, Creation of Plants and Trees/Creation of the Sun, Moon, and Stars. Photo: Kunsthistorisches Institut in Florenz, Max-Planck-Institut fle0009795x_p; Photographer: Roberto Sigismondi.

Plates 

 391

Plate 54. Budapest, Museum of Applied Arts, Salerno ivories, Creation of Birds and Fish. Inv.nr.: 18858. Photo: Ágnes Soltész-Haranghy.

392 

 Plates

Plate 55. New York, Metropolitan Museum of Art, Salerno ivories, Plaque with God Creating the Animals. Accession Number: 17.190.156. (licensed under CC0).

Plates 

 393

Plate 56. Salerno, Museo del Duomo, Salerno ivories, Creation of Eve/Temptation and Fall. Photo: Kunsthistorisches Institut in Florenz, Max-Planck-Institut fle0009801x_p; Photographer: Roberto Sigismondi.

394 

 Plates

Plate 57. Salerno, Museo del Duomo, Salerno ivories, Expulsion/Adam and Eve at Labor. Photo: Kunsthistorisches Institut in Florenz, Max-Planck-Institut fle0009807x_p; Photographer: Roberto Sigismondi.

Plates 

Plate 58. Klagenfurt, Kärntner Landesarchiv, Cod. GV 6/19 (Millstatt Genesis; AT-KLA 118-A-6/19-1 St Genesis), fol. 3v, Forming of Adam. Photo: Kärntner Landesarchiv.

 395

396 

 Plates

Plate 59. Klagenfurt, Kärntner Landesarchiv, Cod. GV 6/19 (Millstatt Genesis; AT-KLA 118-A-6/19-1 St Genesis), fol. 6r, Animation of Adam. Photo: Kärntner Landesarchiv.

Plates 

Plate 60. Klagenfurt, Kärntner Landesarchiv, Cod. GV 6/19 (Millstatt Genesis; AT-KLA 118-A-6/19-1 St Genesis), fol. 8r, Adam in Paradise. Photo: Kärntner Landesarchiv.

 397

398 

 Plates

Plate 61. Klagenfurt, Kärntner Landesarchiv, Cod. GV 6/19 (Millstatt Genesis; AT-KLA 118-A-6/19-1 St Genesis), fol. 9r, Naming of the Animals. Photo: Kärntner Landesarchiv.

Plates 

Plate 62. Klagenfurt, Kärntner Landesarchiv, Cod. GV 6/19 (Millstatt Genesis; AT-KLA 118-A-6/19-1 St Genesis), fol. 9v, Creation of Eve. Photo: Kärntner Landesarchiv.

 399

400 

 Plates

Plate 63. Klagenfurt, Kärntner Landesarchiv, Cod. GV 6/19 (Millstatt Genesis; AT-KLA 118-A-6/19-1 St Genesis), fol. 10r, Temptation of Eve. Photo: Kärntner Landesarchiv.

Plates 

Plate 64. Klagenfurt, Kärntner Landesarchiv, Cod. GV 6/19 (Millstatt Genesis; AT-KLA 118-A-6/19-1 St Genesis), fol. 11r, Fall of Adam and Eve. Photo: Kärntner Landesarchiv.

 401

402 

 Plates

Plate 65. Klagenfurt, Kärntner Landesarchiv, Cod. GV 6/19 (Millstatt Genesis; AT-KLA 118-A-6/19-1 St Genesis), fol. 12r, Reproval of Adam and Eve. Photo: Kärntner Landesarchiv.

Plates 

Plate 66. Klagenfurt, Kärntner Landesarchiv, Cod. GV 6/19 (Millstatt Genesis; AT-KLA 118-A-6/19-1 St Genesis), fol. 14v, Expulsion. Photo: Kärntner Landesarchiv.

 403

404 

 Plates

Plate 67. Klagenfurt, Kärntner Landesarchiv, Cod. GV 6/19 (Millstatt Genesis; AT-KLA 118-A-6/19-1 St Genesis), fol. 16v, Angel Guarding the Gate of Paradise. Photo: Kärntner Landesarchiv.

Plates 

 405

Plate 68. Strasbourg, Bibliothèque de la Ville, Hortus Deliciarum, fol. 17r, Adam and Eve Cycle. Reproduction: Green et al. (1979: fig. 21). Credit: Warburg Institute.

406 

 Plates

Plate 69. Strasbourg, Bibliothèque de la Ville, Hortus Deliciarum, fol. 17v, Adam and Eve Cycle. ­Reproduction: Green et al. (1979: pl. 11). Credit: Warburg Institute.

Plates 

 407

Plate 70. Strasbourg, Bibliothèque de la Ville, Hortus Deliciarum, fol. 27r, Adam and Eve at Labor. Reproduction: Green et al. (1979: pl. 13, no. 24). Credit: Warburg Institute.

408 

 Plates

Plate 71. Oxford, Bodleian Libraries, University of Oxford, Bodleian Library MS. Junius 11, p. 49 (detail), Cain Slaying Abel. Photo: © Bodleian Libraries, University of Oxford (licensed under CC BY-NC 4.0).

Plates 

 409

Plate 72. Venice, Basilica di San Marco. Interior. Mosaics. Atrium. West Arm. South Cupola (The Creation Cupola), Cain Slaying Abel, SM0817. Ekkehard Ritter, North Adriatic Project Fieldwork Records and Papers, 1975–1979, MS.BZ.009, Dumbarton Oaks, Trustees for Harvard University, Washington, D.C.

410 

 Plates

Plate 73. Paris, Bibliothèque Nationale, Ms. lat. 6401 (Boethius), fol. 158v, Christ in Majesty. Source gallica. bnf.fr/Bibliothèque Nationale de France (licensed under CC BY 4.0).

Plates 

Plate 74. Utrecht, Universiteitsbibliotheek, MS Bibl. Rhenotraiectinae I Nr 32 (Utrecht Psalter), fol. 9r (detail), Psalm 17.

 411

412 

 Plates

Plate 75. Utrecht, Universiteitsbibliotheek, MS Bibl. Rhenotraiectinae I Nr 32 (Utrecht Psalter), fol. 26r (detail), Psalm 44.

Plates 

 413

Plate 76. Utrecht, Universiteitsbibliotheek, MS Bibl. Rhenotraiectinae I Nr 32 (Utrecht Psalter), fol. 37v, Psalm 67.

414 

 Plates

Plate 77. Utrecht, Universiteitsbibliotheek, MS Bibl. Rhenotraiectinae I Nr 32 (Utrecht Psalter), fol. 40v (detail), Psalm 71.

Plates 

 415

Plate 78. Utrecht, Universiteitsbibliotheek, MS Bibl. Rhenotraiectinae I Nr 32 (Utrecht Psalter), fol. 41v (detail), Psalm 72.

416 

 Plates

Plate 79. Utrecht, Universiteitsbibliotheek, MS Bibl. Rhenotraiectinae I Nr 32 (Utrecht Psalter), fol. 59v (detail), Psalm 103.

Plates 

 417

Plate 80. Utrecht, Universiteitsbibliotheek, MS Bibl. Rhenotraiectinae I Nr 32 (Utrecht Psalter), fol. 76r (detail), Psalm 134.

418 

 Plates

Plate 81. Utrecht, Universiteitsbibliotheek, MS Bibl. Rhenotraiectinae I Nr 32 (Utrecht Psalter), fol. 82v, Psalm 148.

Plates 

 419

Plate 82. Utrecht, Universiteitsbibliotheek, MS Bibl. Rhenotraiectinae I Nr 32 (Utrecht Psalter), fol. 89v (detail), Canticum 11.

420 

 Plates

Plate 83. Madrid, Biblioteca Nacional de España, MS Vitrina 14-2 (Facundus Beatus), fol. 187r. Images from the collections of the Biblioteca Nacional de España (licensed under CC BY 4.0).

Plates 

 421

Plate 84. Paris, Bibliothèque Nationale, Ms. lat. 8878 (Saint-Sever Beatus), fol. 159r. Source gallica.bnf.fr/ Bibliothèque Nationale de France (licensed under CC BY 4.0).

Seiichi Suzuki The Miniatures and Meters of the Old English Genesis, MS Junius 11

Ergänzungsbände zum Reallexikon der Germanischen Altertumskunde

Herausgegeben von Sebastian Brather, Wilhelm Heizmann und Steffen Patzold

Band 138

Seiichi Suzuki The Miniatures and Meters of the Old English Genesis, MS Junius 11

Volume 2: The Metrical Organization of the Old English Genesis: The Anglo-Saxon Foundations and Old Saxon Adaptation.

ISBN 978-3-11-078688-0 e-ISBN (PDF) 978-3-11-078806-8 e-ISBN (EPUB) 978-3-11-078817-4 ISSN 1866-7678 Library of Congress Control Number: 2022950671 Bibliographic information published by the Deutsche Nationalbibliothek The Deutsche Nationalbibliothek lists this publication in the Deutsche Nationalbibliografie; detailed bibliographic data are available on the internet at http://dnb.dnb.de. © 2023 Walter de Gruyter GmbH, Berlin/Boston Typesetting: Dörlemann Satz GmbH & Co. KG, Lemförde Printing and binding: CPI books GmbH, Leck www.degruyter.com

To my mother, Suzuki Umeko and the memory of my father, Suzuki Keisaku

Preface This book is a companion volume to my concurrent work on the pictures of the Old English Genesis, MS Junius 11 (Volume 1, Suzuki 2023). While addressing independently the metrical and the pictorial (re)organization of the Old English Genesis in their synchronic systematicity and diachronic dynamicity, the two volumes intersect in significant ways, complementing, corroborating, and reinforcing each other in their claims and conjectures. Drawing on my previous work on Old Germanic meter (e.g., Suzuki 1996; 2004; 2014a), this volume provides a formally-oriented, cognitively-based, principled account of the metrical system of the Old English Genesis in its whole network of organization, that is, the meters of Genesis A and Genesis B individually and their coordination and integration at various levels of generalization, from both synchronic–structural and diachronic–dynamic perspectives. Ultimately originating from the Old English and Old Saxon classical meters, the meters of Genesis A and Genesis B were interactively shaped through mutual adaptation and recomposition aimed at their optimal integration into a synthesized Old English Genesis. Among the most significant contributions of the book bearing on the metrical harmonization of Genesis A and Genesis B is the finding that the Old English Genesis is bisected in terms of alliterative pattern between lines 966 and 967. This demonstrated metrical integrity of the earlier portion comprising Genesis A up to line 966 and Genesis B in its entirety, then, may be attributed to its narrative unity. The textual unity thus adduced for explanation is in turn corroborated by the systemic pictorial organization of exclusively the first twenty-two illustrations in MS Junius 11, as substantiated in Volume 1, exactly the set of pictures that are aligned with the part of the text at issue. Furthermore, as argued at length in the preceding volume, these pictures—subsumed under the three cycles, the Fall of the Rebel Angels, the Creation, and Adam and Eve— were originally created at Tours for the thematically corresponding Old Saxon Genesis episodes before their eventual incorporation into the Old English Genesis. The present volume figures prominently in the problematics it defines and addresses, the approaches and methodologies it explores, and the conclusions and conjectures it reaches. First, in previous metrical scholarship, neither Genesis A nor Genesis B has ever received full treatment on its own as an autonomous distinct system, with rare exceptions. Due to its varied deviations from the standard meter, Genesis A is usually dismissed as simply an inexact verse, while, being a translation from Old Saxon, Genesis B is excluded from consideration as an authentic Old English verse. This is a serious scholarly lacuna, because both poems are composed regularly in conformity with their own reorganized meters of Old English, as fully substantiated in this book. Such detailed analysis of the component meters each in turn constitutes an empirical basis for principled account of their interactive reorganization, as explored below. Second, in addition to bringing to light the metrical systematicity of both components of the Old English Genesis, this work is concerned with the emergence of these https://doi.org/10.1515/9783110788068-017

VIII 

 Preface

meters through a close diachronic–comparative examination of the formal and cognitive mechanism by which the Anglo-Saxon poet/editor executed metrical innovations at the junction of the two distinct metrical traditions, Old English and Old Saxon. Third, fully complementary to recent macroscopic works treating a long timespan of metrical evolution (e.g., Weiskott 2016; Cornelius 2017; Russom 2017), this book aims to provide a thorough microscopic analysis of a single poem in its reworking, with a focus on the poet’s active engagement in reorganizing the inherited versification by implementing manifold adaptations and compromises to the heterogeneous and often conflicting conventions he confronted. Fourth, outstanding in methodological terms are the formal rigor and cognitive insight in theorizing, the vigorous and imaginative application of inferential statistics in reasoning, and the close attention to the paradigmatic dimension of the meter—relations based on exclusivity and substitutability, in contrast to the conventional preoccupation with the syntagmatic dimension—relations based on linearity and combinability, that is, individual verse expressions and their internal and external relations. Fifth, the book is multidisciplinary in its research interests and implications. While comparative Germanic linguistics and philology constitute the core as prime disciplines for verse text, the book is constructively fed by and feeds into art historical studies, with due respect for the pictorial dimension of the manuscript illumination involved. Of lasting importance is the conjecture that the Anglo-Saxon poet/translator/editor responsible for the recomposition of the Old English Genesis consulted an illustrated manuscript of Old Saxon Genesis episodes, the hypothesis convergently entailed by the two conclusions independently reached and mutually supporting: that verse lines 1 through 966, composed of Genesis A 1–234, 852–966, and Genesis B—that is, the textual portion corresponding to the pictorial narratives covered by the twenty-two Junius pictures from p. ii to p. 46—are delineated from the rest in terms of alliterative preference; and that these first twenty-two pictures in Junius 11, integrated into a sophisticated system of pictorial organization—the multidimensional network of opposition, complementation, parallelism, and variation in conceptual and compositional terms—are of Touronian origins, produced around the year 850 by a Carolingian artist affiliated with the School of Tours, as substantiated in detail in the first volume (Suzuki 2023). With the rationale for and scope of inquiry formulated above, the book is organized as specified in the Table of Contents and outlined below. Chapter 1, by way of introduction, offers an overview of the Old English Genesis with respect to its textual basis and its unique compositional properties, and presents, from a comparative Germanic perspective, the basics of Old English (Beowulf) and Old Saxon (the Heliand) classical meters—the descriptive and explanatory framework of reference for subsequent exploration—in regard to prototype and its variation along a set of parameters. Chapter 2 deals with the meter of Genesis A. Its earlier, greater part (sections 2.2 through 2.16) is devoted to a comprehensive synchronic account of the meter, as it is transcribed in MS Junius 11, that is, verse lines 1–234 and 852–2936 in the edited text (Doane 2013). The meter of Genesis A thus investigated should be appreciated as an

Preface 

 IX

autonomous system, unified and monolithic in its synchronic functionality (for all its imperfections) rather than a product of inferior versecraft. Subsequently, the later part of this chapter (section 2.17) explores how the original Anglian meter, hypothesized as essentially equivalent to the classical Old English meter, was reconstituted through partial adaptation to the modified Old English meter in the remaking as the Genesis B meter, overlaid with properties derived from Old Saxon versification. Chapter 3 treats the meter of Genesis B. The greater substance of this chapter provides a synchronic account of the meter as a distinct system articulating with Old English meter on the one side and Old Saxon meter on the other (sections 3.2 through 3.16), and uncovers its basically Old English metrical identities while loaded with Old Saxon features modified largely through Anglo-Saxon adaptation (sections 3.17). Specifically, a contrastive investigation into the ways in which the original Old Saxon verses were converted into the Old English ones (section 3.17.1) leads to microscopic analysis of how the meter would likely have undergone reanalysis and reorganization through interlingual recomposition. Of still greater interest for understanding the emergence of the Genesis B meter is to explore whether and how the recipient’s Old English metrical basis maintained its original identity, or conversely had it subverted, in confrontation with influx of extraneous features from the continental source (sections 3.17.2 and 3.17.3). Chapter 4 pursues a wide-ranging organization of the Old English Genesis, now integrated into a single whole, as it crosses the boundaries between the two component meters. Of immediate relevance here is the interpolation of the continental material (Genesis B) in the insular counterpart (Genesis A). Correspondingly, Genesis A is physically—if not as obviously on other grounds—divided into two parts, Genesis A1 (1a–234b) and Genesis A2 (852a–2936b), by the insertion of Genesis B (235a–851b) in between. Such a unique material basis seemingly motivated the separate metrical organization of the two portions. Specifically, there are indications that Genesis A1 and Genesis A2 thus delimited follow separate metrical generalizations in common with or in distinction from Genesis B. On closer inspection, however, the separation of Genesis A1 and Genesis A2 was not directly conditioned by the physical insertion per se of Genesis B: as fully substantiated in section 4.4, and contrary to the usual conceptualization, Genesis A is not divided into two parts—Genesis A1 (1a–234b) and Genesis A2 (852a–2936b)—by the intervening Genesis B as naively conceived. Rather, at odds with the physical interpolation, an alternative more feasible splitting of Genesis A between verses 966b and 967a—that is, at a point well after the interface between Genesis B (851b) and Genesis A (852a)—is evidenced by a significant difference in the probability for double alliteration to occur. Accordingly, the actual bipartition may be characterized as distinguishing the set of Genesis A1/Genesis B (1a–966b) from Genesis A2 (967a–2936b). This overall division of the Old English Genesis stems ultimately from the particular situation in which the underlying Old Saxon Genesis episodes—the antecedent of Genesis B—were available for Old English recomposition, and from the specific ways in which the Old Saxon material was

X 

 Preface

used and concomitantly the narratively overlapping portions of Genesis A were metrically reorganized through partial assimilation to the continental source. Finally, Chapter 5 concludes by reformulating and elaborating on the central claims of the book in cognitively plausible terms and providing conjectures and implications of empirical substance for further study. This work was supported in part by a grant from the Japan Society for the Promotion of Science (JSPS KAKENHI)—a Grant-in-Aid for Scientific Research C (#26370582). The substance of this book was written during the period of this program (2014–2017). Subsequently, during my tenure for 2018–2019 as a Member of the Institute for Advanced Study (IAS), Princeton, I made extensive revisions while working primarily on another project (Volume 1, Suzuki 2023), actually the twin to this book as referred to above. I am deeply grateful to Robbert Dijkgraaf, IAS Director and Leon Levy Professor, Patrick J. Geary, Andrew W. Mellon Professor at IAS, and all staff members at the Institute for support and service they supplied promptly all time. Small parts of this book appeared in print elsewhere: “Metrical positions and their linguistic realisations in Old Germanic metres: A typological overview,” Studia Metrica et Poetica 1, 9–38 (rewritten as sections 1.2 through 1.4); “Three-position verses in Beowulf and Genesis A: Syntagmatically-induced exceptions to the four-position principle,” Journal of Germanic Linguistics 29, 50–84 (incorporated after revision in section 2.13.2). To the editors (Mihhail Lotman and Tracy Alan Hall) and the publishers (University of Tartu Press and Cambridge University Press) of both journals, I wish to express my gratitude for giving permission to use these materials. Worthy of particular thanks is A.  N. Doane, a philologist with a capital P whose exemplary scholarship on and masterful expertise in the Old English Genesis (Doane 1991; 2013) never cease to be awe-inspiring and thought-provoking. Doane (2013: 55n27) kindly mentioned my Genesis meter project still in its infancy, which is at last presented here in completion with my best scholarly regards to him. Finally, it is a pleasant duty to acknowledge, as always, my profound indebtedness to Yoshitaka Tanimoto, former President of Kansai Gaidai University, for his continued support of my research. Seiichi Suzuki May 2020, Neyagawa, Osaka

Contents Preface 

 VII

List of Tables  List of Figures 

 XV  XXI

Symbols and abbreviations 

 XXIII

1

Introduction: The Old English Genesis at the interface of Old English and Old Saxon Meters   1 1.1 The Old English Genesis: Genesis A, Genesis B, and the Old Saxon Genesis   1 1.2 The metrical position as a minimal constituent of a verse, the principle of four positions per verse, and the system of verse types   4 1.3 Variation on the principle of four metrical positions in Old Germanic meter: The verse size   9 1.4 Variation in the alignment of metrical positions with language material   12 1.4.1 The lift   13 1.4.1.1 Resolution   13 1.4.1.2 Suspension of resolution   14 1.4.2 The normal drop   17 1.4.3 The heavy drop   18 1.4.4 Variation in metrical–linguistic alignment: a generalization based on the four-position principle   22

2

 23 The Meter of Genesis A  2.1 Introduction   23 2.2 Type A1 without anacrusis   24 2.3 Type A1 with anacrusis   29 2.4 Subtypes A1s and A2as   34 2.5 Types A2a and A2b   39 2.6 Type A3   43 2.7 Type B1   48 2.8 Type B3   54 2.9 Type C   56 2.10 Types D1, D2a, and D2b   61 2.11 Types D*1, D*2a, and D*2b   64 2.12 Type E   69 2.13 Three-position verses   85

XII 

 Contents

2.13.1 2.13.2 2.14 2.15 2.16 2.17 2.17.1 2.17.2 2.17.3 2.17.4 2.17.5

3

Noncontracted verses   85 Contracted verses   98 The system of verse types   110 Resolution and Kaluza’s law   112 Section-based distinctions   122 Changes introduced under the influence of Old Saxon meter   129 Anacrusis as a full metrical position   130 The reanalysis of PS#px as a reduced variant of PS#Px and the emergence of Px#px   132 Increased preference for -Sx# as a realization of \ × in type E   134 The verse-initial drop in section-initial position   135 Relaxation of the four-position principle in Genesis A as a generalization of adaptations to Old Saxon meter   136

 139 The Meter of Genesis B  3.1 Introduction   139 3.2 Type A1 without anacrusis   140 3.3 Type A1 with anacrusis   144 3.4 Subtype A1s   150 3.5 Types A2a and A2b   150 3.6 Type A3   153 3.7 Type B1   155 3.8 Type B3   160 3.9 Type C   161 3.10 Type D   169 3.11 Type D*   173 3.12 Type E   179 3.13 Three-position verses   184 3.14 The system of verse types   187 3.15 Resolution and Kaluza’s law   189 3.16 Section-based distinctions   198 3.17 The metrical conversion of the Old Saxon Genesis to Genesis B   200 3.17.1 The metrical correspondence between Genesis B and the Old Saxon Genesis   200 3.17.2 Relaxation of the four-position principle for composition of Genesis B as a strategy of conversion and adaptation to Old English meter   211 3.17.3 Metrical features inherited from classical Old English meter, in defiance of the Old Saxon practice   216 3.17.3.1 Prefixal anacrusis   216 3.17.3.2 Alliterative pattern: the relative proportion of single and double alliteration in the a-verse in general   217 3.17.3.3 PS#Px scanned as type A2a   217

Contents 

 XIII

 218 3.17.3.4 The size of the first drop of types A1, A3, B1, and C  3.17.3.5 The distribution of the resolved second lift in type B1   219 3.17.3.6 Lack of distinction between x…PSx/x…Psx and x…PXx/x…Pxx in terms of verse distribution   219 3.17.3.7 A weaker preference for -S- relative to -S# as the first drop of type E   220 3.17.3.8 The frequency of heavy verses   220 3.17.4 The loss of Kaluza’s law   221 4

Split metrical identities of Genesis A: Genesis A1 and Genesis A2 in relation to Genesis B   223 4.1 Introduction   223 4.2 Alliteration: increased implementation of double alliteration in Genesis A1 in concert with Genesis B and in contrast with Genesis A2   224 4.3 Further distinguishing properties   229 4.3.1 Size of anacrusis   229 4.3.2 The initial upbeat of type A3   230 4.3.3 The extra drop of class D*   230 4.3.4 The first drop of type E: -S# versus -S-   231 4.3.5 Subtypes A1s/A2as relative to types A1/A2a   232 4.3.6 Summary   232 4.4 The pictorial and textual foundations of Genesis A1: assimilation to Genesis B and dissimilation to Genesis A2   234 4.5 The mechanism of boosting double alliteration in Genesis A1   253

5 Conclusion  Bibliography 

 267  283

 287 Index  Index of authors   287 Index of subjects   288 Index of verses cited for discussion or exemplification 

 292

List of Tables Table 1.1. Table 1.2. Table 1.3.

Formation of basic verse types   7 The system of verse types   8 Occurrences of anacrustic and nonanacrustic configurations (classes A, D, and D*, excluding type A3)   10 Table 1.4. Occurrences of class D* relative to class D (excluding anacrusis)   11 Table 1.5. Occurrences of anacrusis in class D* (all variants)   11 Table 1.6. Resolution on the first lift   14 Table 1.7. Resolution on the second lift   14 Table 1.8. Occurrences of PS#px, Pxpx, and Px#Px   15 Table 1.9. Occurrences of x…Ppx and x…PPx   15 Table 1.10. Size of the verse-initial drop in types B1 and C in terms of syllable numbers realized   17 Table 1.11. Size of the verse-final drop of type A1 in terms of syllable numbers realized   18 Table 1.12. Realizations of the heavy drop of type E according to syllable shape   21 Table 2.1. Table 2.2. Table 2.3. Table 2.4. Table 2.5. Table 2.6. Table 2.7. Table 2.8. Table 2.9. Table 2.10. Table 2.11. Table 2.12. Table 2.13. Table 2.14. Table 2.15. Table 2.16. Table 2.17. Table 2.18. Table 2.19. Table 2.20. Table 2.21. Table 2.22.

Verse distribution and alliterative pattern of type A1 according to the size of the first drop in Genesis A (excluding anacrusis)   24 Size of the first drop of type A1 in Genesis A and Beowulf   25 Verse distribution and alliterative pattern of type A1 according to the variation of the monosyllabic first drop in Genesis A   26 Verse distribution and alliterative pattern of type A1 according to the variation of the disyllabic first drop in Genesis A   27 Verse distribution of PxSx in Genesis A, Beowulf, and the Heliand   28 Anacrusis and verse distribution of type A1 in Genesis A, Beowulf, and the Heliand   29 Anacrusis and alliterative pattern of type A1 in the a-verse in Genesis A, Beowulf, and the Heliand   30 Occurrence of anacrusis in type A1 according to the size of the first drop in Genesis A, Beowulf, and the Heliand   31 Occurrence of anacrusis in type A1 according to the variation of the monosyllabic first drop in Genesis A, Beowulf, and the Heliand   32 Size of anacrusis in type A1 in Genesis A and Beowulf   33 Verse distribution and alliterative pattern of variants of subtypes A1s/A2as (with the monosyllabic first drop) in Genesis A and Beowulf   35 Verse distribution and alliterative pattern of PS#Px and PS(#)x(#)…Px in Genesis A   40 Resolution on the heavy drop of type A2a in Genesis A and Beowulf   41 Resolution on the first lift of type A2a in Genesis A and Beowulf   41 Verse distribution and alliterative pattern of Px…PS, PS#PS, and Px…PP in Genesis A   42 Verse distribution of type A2b (Px…PS) according to the size of the first drop in Genesis A   43 Verse distribution of type A2b (PxPS) according to the variation of the monosyllabic first drop in Genesis A   43 Size of the initial upbeat in type A3 in Genesis A and Beowulf   46 Verse distribution of type A3 variants in Genesis A and Beowulf   48 Verse distribution and alliterative pattern of type B1 variants in Genesis A   49 Verse distribution of type B1 according to the size of the first drop in Genesis A and Beowulf   50 Alliterative pattern of type B1 according to the size of the first drop in Genesis A and Beowulf (excluding x…PxS)   51

https://doi.org/10.1515/9783110788068-018

XVI 

 List of Tables

Table 2.23. Verse distribution of type B1 according to the size of the second drop in Genesis A and Beowulf   51 Table 2.24. Alliterative pattern of type B1 according to the size of the second drop in Genesis A and Beowulf (excluding x…PxS)   52 Table 2.25. Verse distribution of type B1 according to the variation of the monosyllabic second drop in Genesis A and Beowulf   53 Table 2.26. Alliterative pattern of type B1 according to the variation of the monosyllabic second drop in Genesis A and Beowulf (excluding -x-)   53 Table 2.27. Verse distribution of the long and the resolved second lift in type B1 in Genesis A and Beowulf   53 Table 2.28. Verse distribution of the long and the resolved lift in type B3 in Genesis A   55 Table 2.29. Verse distribution and alliterative pattern of type C variants in Genesis A and Beowulf   57 Table 2.30. Verse distribution of type C according to the size of the first drop in Genesis A and Beowulf   57 Table 2.31. Alliterative pattern of type C according to the size of the first drop in Genesis A and Beowulf (with exclusive reference to variants with the primary-stressed second lift)   58 Table 2.32. Size of the first drop of type C according to the stress variation of the second lift in Genesis A and Beowulf   59 Table 2.33. Size of the first drop and length of the first lift of type C in Genesis A and Beowulf   59 Table 2.34. Size of the first drop and length of the second lift of type C lift in Genesis A and Beowulf   60 Table 2.35. Variation of the first and second lifts of type C in Genesis A and Beowulf   61 Table 2.36. Verse distribution and alliterative pattern of types D1, D2a, and D2b in Genesis A and Beowulf   62 Table 2.37. Anacrusis and verse distribution of class D in Genesis A and Beowulf   63 Table 2.38. Occurrence of anacrusis in types D1, D2a, and D2b in Genesis A and Beowulf   64 Table 2.39. Verse distribution and alliterative pattern of types D*1, D*2a, and D*2b in Genesis A (including five instances with anacrusis) and Beowulf   67 Table 2.40. Realizations of the extra drop of types D*1, D*2a, and D*2b in Genesis A and Beowulf   68 Table 2.41. Verse distribution and alliterative pattern of type E variants ending in P in Genesis A   72 Table 2.42. Verse distribution and alliterative pattern of the configurations ending in px in contrast to those in Px in Genesis A   77 Table 2.43. Categorization of configurations ending in P/px as types A1, A2a, or E in Genesis A and Beowulf   79 Table 2.44. The system of verse types in Genesis A   111 Table 2.45. Occurrence of short and long disyllables according to metrical context in Genesis A and Beowulf   121 Table 2.46. Size of sections in terms of the number of constituent verses in Genesis A, Beowulf, and the Heliand   124 Table 2.47. Proportion of lift-initial to drop-initial verses in Genesis A, Genesis B, Beowulf, and the Heliand   126 Table 2.48. Size of the verse-initial upbeat in the section-initial versus noninitial a-verses of Genesis A, Beowulf, and the Heliand   129 Table 2.49. Verse distribution of PS#px, PS#Px, and Px#Px in Genesis A, Beowulf, and the Heliand   133 Table 2.50. Categorization of PS#px, PS#Px, and Px#Px in Genesis A, Beowulf, and the Heliand on the basis of their verse distribution measured in p-values by Fisher’s test   133 Table 3.1.

Verse distribution and alliterative pattern of type A1 according to the size of the first drop in Genesis B (excluding anacrusis)   141

List of Tables 

 XVII

Table 3.2. Size of the first drop of type A1 in Genesis B, Genesis A, Beowulf, the Heliand, and the Old Saxon Genesis   142 Table 3.3. Verse distribution and alliterative pattern of type A1 according to the variation of the monosyllabic first drop in Genesis B   142 Table 3.4. Verse distribution of PxSx in Genesis B, Genesis A, Beowulf, the Heliand, and the Old Saxon Genesis   143 Table 3.5. Anacrusis and verse distribution of type A1 in Genesis B, Genesis A, Beowulf, the Heliand, and the Old Saxon Genesis   144 Table 3.6. Anacrusis and alliterative pattern of type A1 in Genesis B, Genesis A, Beowulf, the Heliand, and the Old Saxon Genesis   145 Table 3.7. Occurrence of anacrusis in type A1 according to the size of the first drop in Genesis B, Genesis A, Beowulf, the Heliand, and the Old Saxon Genesis   146 Table 3.8. Occurrence of anacrusis in type A1 according to the variation of the monosyllabic first drop in Genesis B, Genesis A, Beowulf, the Heliand, and the Old Saxon Genesis   147 Table 3.9. Size of anacrusis in type A1 in Genesis B, Genesis A, Beowulf, the Heliand, and the Old Saxon Genesis   149 Table 3.10. Verse distribution and alliterative pattern of PS#Px, PSx#Px, Px…PS, and Px…Ppx/Px…Psx in Genesis B   151 Table 3.11. Size of the initial upbeat in type A3 in Genesis B, Genesis A, Beowulf, the Heliand, and the Old Saxon Genesis   154 Table 3.12. Size of the first drop of type B1 in Genesis B, Genesis A, Beowulf, the Heliand, and the Old Saxon Genesis   155 Table 3.13. Verse distribution and alliterative pattern of type B1 according to the size of the first drop in Genesis B   156 Table 3.14. Verse distribution and alliterative pattern of type B1 according to the size of the second drop in Genesis B, the Heliand, and the Old Saxon Genesis (excluding x…PxS for alliterative pattern in the a-verse)   158 Table 3.15. Verse distribution of type B1 according to the variation of the monosyllabic second drop in Genesis B, the Heliand, and the Old Saxon Genesis   159 Table 3.16. Alliterative pattern of type B1 according to the variation of the monosyllabic second drop in Genesis B, the Heliand, and the Old Saxon Genesis (excluding -x- due to its inherent bias toward single alliteration)   159 Table 3.17. Verse distribution of the long and the resolved second lift in type B1 in Genesis B, the Heliand, and the Old Saxon Genesis   160 Table 3.18. Verse distribution and alliterative pattern of type C variants in Genesis B, the Heliand, and the Old Saxon Genesis   162 Table 3.19. Size of the first drop of type C in Genesis B, Genesis A, Beowulf, the Heliand, and the Old Saxon Genesis   163 Table 3.20. Verse distribution and alliterative pattern of type C according to the size of the first drop in Genesis B, the Heliand, and the Old Saxon Genesis   164 Table 3.21. Size of the first drop of type C according to the stress variation of the second lift in Genesis B, the Heliand, and the Old Saxon Genesis   166 Table 3.22. Size of the first drop and length of the first lift in type C in Genesis B, the Heliand, and the Old Saxon Genesis   167 Table 3.23. Size of the first drop and length of the second lift in type C in Genesis B, the Heliand, and the Old Saxon Genesis   168 Table 3.24. Variation of the first and second lifts of type C in Genesis B, the Heliand, and the Old Saxon Genesis   169

XVIII 

 List of Tables

Table 3.25. Verse distribution and alliterative pattern of D1, D2a, and D2b in Genesis B, the Heliand, and the Old Saxon Genesis (excluding anacrusis)   170 Table 3.26. Anacrusis and verse distribution of class D in Genesis B, the Heliand, and the Old Saxon Genesis   172 Table 3.27. Occurrence of anacrusis in D1, D2a, and D2b in Genesis B, the Heliand, and the Old Saxon Genesis   172 Table 3.28. Size of anacrusis in class D in Genesis B, Genesis A, Beowulf, the Heliand, and the Old Saxon Genesis   173 Table 3.29. Verse distribution and alliterative pattern of D*1, D*2a, and D*2b in Genesis B, the Heliand, and the Old Saxon Genesis (excluding anacrusis)   174 Table 3.30. Proportion of D1, D2a, and D2b to D*1, D*2a, and D*2b in Genesis B, Genesis A, Beowulf, the Heliand, and the Old Saxon Genesis   175 Table 3.31. Anacrusis and verse distribution of class D* in Genesis B, Genesis A, Beowulf, the Heliand, and the Old Saxon Genesis   176 Table 3.32. Occurrence of anacrusis in D*1, D*2a, and D*2b in Genesis B, Genesis A, Beowulf, the Heliand, and the Old Saxon Genesis   177 Table 3.33. Size of anacrusis in class D* in Genesis B, Genesis A, Beowulf, the Heliand, and the Old Saxon Genesis   178 Table 3.34. Realizations of the extra drop of class D* in Genesis B (including anacrusis)   178 Table 3.35. Verse distribution and alliterative pattern of type E variants ending in P in Genesis B (without anacrusis)   179 Table 3.36. Categorization of x…PSx…P and x…PXx…P as type E with anacrusis and/or type B1 in Genesis B, Genesis A, Beowulf, the Heliand, and the Old Saxon Genesis   183 Table 3.37. Statistical difference in terms of p-values in the probability of three-position verses in Genesis B, Genesis A, Beowulf, the Heliand, and the Old Saxon Genesis   185 Table 3.38. Statistical difference in terms of p-values in the ratio of the long and resolved first lifts in three-position verses in Genesis B, Genesis A, Beowulf, the Heliand, and the Old Saxon Genesis   186 Table 3.39. The system of verse types in Genesis B   187 Table 3.40. Occurrence of short and long disyllables according to metrical context in Genesis B and Beowulf   192 Table 3.41. Occurrence of short and long disyllables according to metrical context in Genesis B and  193 Genesis A  Table 3.42. Occurrence of short and long disyllables according to metrical context in Genesis B and the Old Saxon Genesis   196 Table 3.43. Occurrence of open and closed disyllables according to metrical context in Genesis B and the Old Saxon Genesis   197 Table 3.44. Verse distribution of type A1 according to the size of anacrusis in Genesis B, the Heliand, and the Old Saxon Genesis   204 Table 3.45. Correlation of the number of verse-initial unstressed syllables to the a-/b-verse distinction in Genesis B, Genesis A, Beowulf, the Heliand, and the Old Saxon Genesis   204 Table 4.1. Table 4.2. Table 4.3. Table 4.4. Table 4.5.

Single versus double alliteration in the a-verse in Genesis A1, Genesis B, and Genesis A2   224 Size of the first drop of type A1 (including anacrusis) in Genesis A1 and A2   226 Size of anacrusis in terms of syllable numbers in Genesis A1, Genesis B, and Genesis A2   230 Size of the initial upbeat of type A3 in terms of syllable numbers in Genesis A1, Genesis B, and Genesis A2   230 Frequency of PSx#…P and PS#x…P in Genesis A1, Genesis B, and Genesis A2   232

List of Tables 

 XIX

Table 4.6. Frequency of types A1 and A2a and subtypes A1s/A2as in Genesis A1, Genesis B, and Genesis A2   232 Table 4.7. Statistical differences in terms of p-values between Genesis A1 and Genesis B along the six metrical parameters   235 Table 4.8. Statistical differences in terms of p-values between Genesis A1 and Genesis A2 along the six metrical parameters   235 Table 4.9. Statistical differences in terms of p-values between Genesis B and Genesis A2 along the six metrical parameters   235 Table 4.10. Single versus double alliteration in the a-verse in Genesis A1, Genesis B, and Genesis A2 according to the varying boundary between Genesis A1 and Genesis A2   236 Table 4.11. Size of anacrusis in Genesis A1, Genesis B, and Genesis A2 according to the varying boundary between Genesis A1 and Genesis A2   236 Table 4.12. Size of the initial upbeat of type A3 in Genesis A1, Genesis B, and Genesis A2 according to the varying boundary between Genesis A1 and Genesis A2   237 Table 4.13. Size of the extra drop of class D* in Genesis A1, Genesis B, and Genesis A2 according to the varying boundary between Genesis A1 and Genesis A2   237 Table 4.14. Frequency of PSx#…P and PS#x…P in Genesis A1, Genesis B, and Genesis A2 according to the varying boundary between Genesis A1 and Genesis A2   238 Table 4.15. Frequency of types A1/A2a and subtypes A1s/A2as in Genesis A1, Genesis B, and Genesis A2 according to the varying boundary between Genesis A1 and Genesis A2   238 Table 4.16. Frequency of double alliteration per 100 a-verses according to sections in the Old English Genesis   243 Table 4.17. Parameter estimates report for the regression coefficient on S in Beowulf, the Heliand, Genesis A, and Genesis B   245 Table 4.18. Parameter estimates report for the regression coefficient on S in Genesis A1/Genesis B and Genesis A2 according to their four varying boundaries   247 Table 4.19. Alliterative pattern of verse classes in the a-verse in Genesis A1, Genesis B, and Genesis A2   254 Table 4.20. Distribution of verse classes in Genesis A1, Genesis B, and Genesis A2   254 Table 4.21. Single versus double alliteration in type A2 in Genesis A1, Genesis B, and Genesis A2   255 Table 4.22. Single versus double alliteration in D*2a and D*2b in Genesis A1, Genesis B, and Genesis A2   256 Table 4.23. Single versus double alliteration in type A1 (Px…Px) in Genesis A1, Genesis B, and Genesis A2 (excluding anacrusis, PxSx, and subtype A1s)   257 Table 4.24. Single versus double alliteration in type B1 (x…Px…P) in Genesis A1, Genesis B, and Genesis A2   257 Table 4.25. Single versus double alliteration in type C (x…PPx) in Genesis A1, Genesis B, and Genesis A2   258 Table 4.26. Single versus double alliteration in D1 (PPxx) in Genesis A1, Genesis B, and Genesis A2 (excluding anacrusis and PSxx)   258 Table 4.27. Single versus double alliteration in D*1 (PxPxx) in Genesis A1, Genesis B, and Genesis A2 (excluding anacrusis)   258 Table 4.28. Single versus double alliteration in type E in general in Genesis A1, Genesis B, and Genesis A2 (excluding anacrusis)   259 Table 4.29. Size of the first drop of type A1 in Genesis A1, Genesis B, and Genesis A2 (excluding PxSx and anacrusis)   260 Table 4.30. Single versus double alliteration in type A1 variants with the monosyllabic first drop (Px#Px/P#xPx/P#x#Px) in Genesis A1, Genesis B, and Genesis A2   260 Table 4.31. Single versus double alliteration in Px#Px in Genesis A1, Genesis B, and Genesis A2   262

XX 

 List of Tables

Table 4.32. Single versus double alliteration in P#xPx and P#x#Px in Genesis A1, Genesis B, and Genesis A2   262 Table 4.33. Single versus double alliteration in type A1 with the polysyllabic first drop in Genesis A1, Genesis B, and Genesis A2   262 Table 4.34. Single versus double alliteration in type B1 according to the size of the first drop in Genesis A1, Genesis B, and Genesis A2   263 Table 4.35. Single versus double alliteration in type C according to the size of the first drop in Genesis A1, Genesis B, and Genesis A2   264 Table 5.1. A metrical stratification of the Old English Genesis   268 Table 5.2. A simplex metrical stratification of the Old English Genesis (corresponding to a conventional conceptualization)   268

List of Figures Figure 4.1. Bivariate analysis of sections and double alliteration in Beowulf   241 Figure 4.2. Bivariate analysis of sections and double alliteration in the Heliand   242 Figure 4.3. Bivariate analysis of sections and double alliteration in Genesis B   244 Figure 4.4. Bivariate analysis of sections and double alliteration in Genesis A   245

https://doi.org/10.1515/9783110788068-019

Symbols and abbreviations a alliterating lift in contrast to ‘x’ (nonalliterating lift) acc. accusative adj. adjective adv. adverb anom. anomalous verb Beo Beowulf C consonant dat. dative fem. feminine GenA Genesis A GenB Genesis B Go. Gothic Hel Heliand ind. indicative neut. neuter nom. nominative OE Old English ON Old Norse OS Old Saxon OSG Old Saxon Genesis P primary-stressed long syllable p primary-stressed short syllable PGmc. Proto-Germanic pl. plural pres. present pret. preterite pron. pronoun S secondary-stressed long syllable s secondary-stressed short syllable sg. singular st. strong verb subj. subjunctive V vowel wk. weak verb X unstressed long syllable x unstressed short syllable or unstressed syllable in general; nonalliterating lift in contrast to ‘a’ (alliterating lift) x… one or more unstressed syllables # word boundary / lift \ heavy drop × normal drop > becomes; ranks higher than < derives from; ranks lower than ≥ apparently ranks higher than, but lacking statistical support

https://doi.org/10.1515/9783110788068-020

1 Introduction: The Old English Genesis at the interface of Old English and Old Saxon Meters 1.1 The Old English Genesis: Genesis A, Genesis B, and the Old Saxon Genesis The Old English Genesis—often referred to as the Caedmon(ian) Genesis owing to the misplaced attribution of its authorship to Caedmon in earlier scholarship—is a biblical poem composed in accordance with the Anglo-Saxon alliterative tradition. The work is located in an illustrated manuscript, Oxford, Bodleian Library, MS Junius 11 (Gollancz 1927; Krapp 1931; Ker 1957: 406–408; Temple 1976: 76–78; Lucas 1980; 1981; Raw 1984; Karkov 2001: 19–32; Lockett 2002; Doane 2013: 1–41; Gneuss/Lapidge 2014: 491–493; Bodleian Library MS. Junius 11). Containing 2936 verse lines, the poem constitutes the longest Old English verse text after Beowulf (3182 lines; Fulk/Bjork/Niles 2008). The poem is divided into two parts corresponding to its double origins in terms of composition: Genesis A (Doane 2013; verse lines 1–234 and verse lines 852–2936; pp. 1–12/18 and pp. 40/8–142/8; key: page/line in the manuscript) and Genesis B (Doane 1991; verse lines 235–851; pp. 13–40/8). On the one hand, Genesis A is based on an insular original composed in Old English, broadly in eighth century Mercia or Northumbria (cf. Cable 1981: 80–81; Fulk 1992: 60–65, 348–351, 392; Cronan 2004: 47–49; Doane 2013: 51–55; Neidorf 2013–2014: 34–40). This antecedent Old English work underlying Genesis A may be referred to as the Anglian Genesis (section 2.1). On the other hand, Genesis B is an Old English translation of some episodes constitutive of the Old Saxon alliterative biblical verse text, composed on the continent in the mid-ninth century (ca. 830s; Doane 1991; see also below) and reflected in the Old Saxon Genesis fragments (Doane 1991: 232–252; Behaghel/Taeger 1996: 241–256; Vatican, Biblioteca Apostolica, Palatinus Latinus 1447, fols. 1r, 2r, 2v, and 10v). Thus, the extant Old English Genesis is an amalgamation of the two thematically related works of distinct compositional identities by embedding the translated continental material (resulting in Genesis B) into the native work (resulting in Genesis A). Yet the resultant work would have been intended as a unitary piece, for all the obvious differences involved between the two constituents in matters of language, style, meter, diction, and most importantly, the ways in which the biblical materials are used for poetic narrative. Uniquely embodying interlingual conversions of verbal and visual representations of Genesis episodes, the Old English Genesis was an ambitious experimental project, and as such it invites intensive cross-disciplinary investigation in both the synchronic and diachronic dimensions of its multifaceted reworking. The designed integrity of the two disparate works into a single whole is confidently ascertained in the absence of any explicit marking of the break between the two constituents. Specifically, there are no indications in evidence of a textual discontinuity on https://doi.org/10.1515/9783110788068-021

2 

 1 Introduction

the eighth line of page 40 (p. 40/8) in the extant manuscript in which Genesis B reverts to Genesis A at a transition from Genesis B 851b to Genesis A 852a. The total lack of a formal distinction seems significant, given that a variety of marking devices would have been independently available to the scribe for appropriation: a capitalization and illumination of a section-initial letter, a cluster of dots for signaling a section ending, a picture inserted at a section boundary, or presumably the simplest means—leaving a blank space after verse 851b and starting verse 852a at the following line (i.e., p. 40/9).1 The translation and interpolation of the Old Saxon source obviously determined the identity of the resultant Old English Genesis no less decidedly than the underlying native work that framed the foreign material. In light of the demonstrable integrity of the final product characterized by the smooth transition from Genesis B to Genesis A observed above, it may be warranted to infer that the poet in charge executed other comparable editorial practices for harmonization and integration of the two constituents. To situate this stage of poetic merger more specifically in the whole evolution of the Old English Genesis, one may postulate the following series of events in the early West Saxon period, in agreement with Doane: About 900 the direct ancestor of Genesis A was combined with Genesis B. The first part of Genesis A received a revision introducing many early West-Saxon forms in the process of being combined with the recently or simultaneously west-saxonized Genesis B. (Doane 2013: 40)

Among the many conceivable domains targeted for revision and refinement would no doubt have been the metrical organization, as the meter is the prime defining basis of verse. Moreover, it is not only Genesis B that underwent metrical revision by way of Anglo-Saxonization as an Old English poem; Genesis A also would have been implicated no less significantly in editorial reorganization and integration to effect a better harmonization with Genesis B. That is, one must conceptualize that, with a view to attaining a unity of the whole work, the editorial reworking would have been performed through negotiation and compromise between the Old English and Old Saxon meters. On the one side, the meter of the immediate predecessor of Genesis A—the Anglian Genesis—would have been reorganized under the influence of the Old Saxon meter through the intermediary of the Old English translation of the relevant portions of the Old Saxon Genesis. On the other side, the meter of Genesis B would have come into being through restructuring of its pre-edited version, largely in concert with the surrounding Old English material (i.e., the pre-edited version of Genesis A). Thus, situated at the interface of the insular and continental traditions, the meters of Genesis A and Genesis B would eventually have emerged in their extant form through mutual adaptation and reciprocal reorganization as integral constituents of the Old English Genesis at large.

1 The exact place of transition from Genesis A to Genesis B is unrecoverable, because two leaves were lost between the end of Genesis A (234b; p. 12/18) and the beginning of Genesis B (235a; p. 13/1) in the current manuscript (Doane 2013: 11). For details, see section 4.1 below, and also section 1.1, Volume 1.

1.1 The Old English Genesis: Genesis A, Genesis B, and the Old Saxon Genesis 

 3

The ultimate objectives of this book, therefore, are to substantiate the ways in which and the means by which such metrical revision and refinement would have been exercised and to evaluate accomplishments brought about (with varying degrees of success) by the strategy of metrical integration. Since, methodologically and epistemologically, static–synchronic investigations of the resultant, directly observable meters precede dynamic–diachronic ones addressing their earlier (i.e., premanuscript) stages, one must prioritize exploration into the attested meters of Genesis A and Genesis B on their own terms as autonomous systems. Yet as they are embedded in the metrical traditions, they need to be appreciated in relation to the traditional canons. Under the usual working hypothesis that the Old English and Old Saxon meters are fully exemplified in Beowulf (Fulk/Bjork/Niles 2008) and the Heliand (Behaghel/Taeger 1996), respectively, it may be justified to use these classical works as the overall metrical framework, by reference to and in comparison with which the synchrony of the two Old English Genesis meters (Genesis A and B) will be subject to systematic description and made to yield viable dataset for a contact-based evolutionary account to be pursued from a dynamic–diachronic perspective. In more specific terms, the major deviations from the canonical meters will be substantiated to have arisen through mutual adaptation and reorganization implemented as part of the strategy for integration of the Old English Genesis into a single text. With the rationale and problematics for exploration thus defined, the remainder of this introductory chapter provides a contrastive account of Old Germanic meter by illustrating its parametric variations that largely determine the metrical identities of the two major West Germanic alliterative poems, the Old English Beowulf and the Old Saxon Heliand.2 The primary parameters to be examined here are the principle of four metrical positions per verse and its varying binding force (sections 1.2 and 1.3), and the differing ways in which these constituent positions are aligned with linguistic material (metrical–linguistic alignment; section 1.4). On the one hand, the four-position principle works with a maximal strictness in Beowulf, whereas it allows for a wider range of deviations in verse size in the Heliand. On the other hand, the variation in the metrical– linguistic alignment in the two close cognate meters may be generalized by positing the common scale, Heliand > Beowulf, whereby the Heliand is distinguished in overall terms by the higher likelihood of resolution, the lower likelihood of suspending resolution, and the larger size of the drop, all being ultimately reducible to the greater metrical prominence inherent in the Old Saxon poem.

2 The following sections are an abridged and condensed version of Suzuki (2014b). For a fuller survey taking the Norse meters into account, see Suzuki (2014a; 2014b).

4 

 1 Introduction

1.2 The metrical position as a minimal constituent of a verse, the principle of four positions per verse, and the system of verse types Drawing on Suzuki (1996: 10, 371; 2004: 7; 2014a: 13–15), I conceptualize the meter as a prototype-based cognitive system of rules, constraints, and representations. Such theoretical constructs are categorized and computed in terms of the graded opposition between their prototypes (core/central/unmarked members) and derivative variants (peripheral/outlying/marked members). Predicated thus on prototype in categorization, the system legitimates a given verbal expression as a verse (or a metrical unit sub- or superordinate to it; see below) by determining its metricality in gradient terms along a set of relevant parameters or preference conditions. Since the rules and constraints postulated to account for metrical phenomena are formulated and executed differentially, corresponding to varying prototypicality, and thus allow for a range of gradable deviations in a principled way, the metrical system is inherently subject to parametrically patterned variations, which are amenable to frequency-based descriptions and stochastic generalizations. The meter thus conceptualized organizes versification at multiple levels according to the hierarchy of metrical units. The significant units in Old Germanic meter include the following (from bottom to top in the metrical hierarchy): position; verse; line; and section. The metrical position, the lowest in the prosodic hierarchy, is postulated as an abstract invariant minimal metrical unit underlying a host of linguistic realizations that vary widely along a wealth of parameters including stress, syllable length and number, and lexical properties. Equipped with this underlying unit, one is principally concerned with the differing ways in which the metrical position is aligned with various language materials by versification. In other words, the central issues of Old Germanic metrics are to account for the range and likelihood of legitimate linguistic realizations of a given metrical position according to context. Thus, the varying alignments of metrical positions with language materials serve as significant parameters along which individual cognate meters differ in the evolution of Old Germanic versification (section 1.4). It was Cable (1974: 84–93) among other metrists who brilliantly brought to light the primacy of metrical positions (or members) as foundational units of Old English alliterative verse: he laid out the arrangement of four metrical positions as the fundamental principle of verse composition in Old English. Indeed, Sievers (1893: 25) had earlier rightfully conceptualized this unit (Glieder) as the smallest metrical constituent, but he failed to fully work out its implications to the logical conclusion (Cable 1974: 32), which Cable accomplished with admirable clarity. By addressing the grouping of these smallest constituents into feet, Sievers obscured rather than clarified their primary status in the meter (Cable 1974: 91). Thus, Cable (1974) firmly laid the foundation for Old English (and Old Germanic in general by implication) metrics by his articulate formulation of the

1.2 The metrical position as a minimal constituent of a verse 

 5

principle of four positions as its ultimate basis of versification—a verse is composed of four metrical positions. The four-position principle formulated above differs in its exact manners of execution from one meter to another in the Old Germanic alliterative traditions: it prescribes versification with varying strictness and flexibility. The principle of four positions is accordingly involved in parametric variations along a number of attributes, thereby accounting for the variability of Old Germanic meters in no small measure (section 1.3). Apart from the canonical number of constituent positions dictated, their identities and arrangement are specifically prescribed. Building on a binary opposition between lift (strong position, represented as /) and drop (weak position, represented as ×), each verse normally comprises two such lifts and drops, respectively. A further distinction—a subcategorization of the drop—may be made between normal (×) and heavy (\) drops through promotion of a (normal) drop to a relatively prominent variant, a heavy drop. A heavy drop is substituted for a normal one by a secondary operation of metrical strengthening (or promotion). The set thus organized of the binary-opposed metrical positions—lift versus drop—supplemented with a further two-way subdivision—normal versus heavy drops—is differentiated optimally according to their prototypical metrical–linguistic alignment. The lift is aligned with a primary-stressed long syllable (P); the heavy drop, with a secondary-stressed long syllable (S); and the normal drop, with an unstressed syllable or a string thereof (x…).3 The arrangement of these lifts and drops in turn is subject to constraint (Suzuki 1996: 18–19). A concatenation of two (normal) drops in nonfinal position is disallowed. This prohibition stems from the generalization just presented that a nonfinal (normal) drop may be occupied by a varying quantity of linguistic material, to be determined in due course. Since a nonfinal drop is thus qualified for alignment with multiple syllables in its realization, the nonfinal succession of two drops × × is formally equivalent to the occurrence of a single drop × in terms of linguistic materialization. This reducibility of two successive drops to one gives rise to the following regulations as its derivatives. First, at least one of the first two positions must be strong. That is, the concatenation × × / / is categorically excluded, because it is indistinguishable from the illegitimate configuration *× / /, an ill-formed, three-position construct that violates the four-position requirement. Second, and by the same token, the sequence / × × / is disallowed, as it is counted as equal to the sequence / × /, too short a verse comprising only three positions. By contrast, the configuration ending in × ×, namely, / / × ×, is licensed because of the strictly monosyllabic realization of the verse-final drop. Due to the invariable monosyllabicity, the occurrence of more than one unstressed syllable at the end of a verse indicates unequivocally the presence of two successive normal drops. Unlike the unmetrical verse-initial sequence × ×, the verse-internal counterpart is susceptible to improvement so as to be admitted eventually into the set of legitimate

3 For marked modes of alignment, see section 1.4 below.

6 

 1 Introduction

configurations (verse classes and basic verse types; see below). Specifically, in place of a succession of normal drops that is no different in effect from a single drop in terms of linguistic manifestation, the first one is promoted to a heavy variant \ by metrical strengthening (to be treated fully in due course), thereby leading to the configuration / \ × /, classified as type E (see below). Two questions arise at this point. First, why was / × \ / not allowed to occur as an improved alternative? Second, why was the verse-initial sequence × × immune to a comparable drop promotion creating *\ × / /? The first question is adequately addressed by invoking the linearity-based prominence scale of metrical positions, whereby, of two like positions, the one closer to the beginning of a verse is identified as more prominent than the other (Suzuki 1996: 167; 2004: 10; 2014a: 11). Accordingly, given the string × ×, the first one is characterized as more prominent than the second. Inasmuch as a promotion to a heavy drop increases prominence, strengthening the first drop—the inherently more salient one—may be viewed as the more natural execution than the alternative of adding prominence to the second relatively weaker one at variance with the general contour of falling prominence. The second question is subject to a principled account based on linguistic structure. At stake is the prototypical realization of a heavy drop. It is prototypically embodied by a secondary-stressed syllable, as noted above. In Old Germanic versification, the verse-initial position invariably matches the word-initial position, as no verse is allowed to occur that begins with a medial or word-final syllable. Since, with rare exceptions, a secondary-stressed syllable normally appears after a primary-stressed one, with or without an intervening unstressed syllable, the concatenation \ (×) / virtually lacks a linguistic reality. There is therefore no way of constructing a verse starting with a secondary-stressed syllable. Accordingly, metrical strengthening is precluded from implementation on the verse-initial drop. In summary, the principle of four metrical positions and their rule-governed arrangement determined on the basis of the primary opposition between lift and (normal) drop and their prototypical metrical–linguistic alignment define the following inventory of configurations as basic verse types in Old Germanic poetry: (1)          

Five basic verse types Type A: / × / × Type B: × / × / Type C: × / / × Type D: / / × × Type E: / \ × /

Following conventional practice (e.g., Sievers 1893), the alphabetical labels given to the five basic types A through E are designed to correspond in descending order with their relative frequency, with type A / × / × occurring with maximal incidence. As shown below, the classificatory labels will be elaborated as derived verse types are added to

1.2 The metrical position as a minimal constituent of a verse 

 7

the system. Accordingly, designations A through E without subdividing markers will be reused to refer to verse classes, superordinate to verse types. The static manner of representation as in (1) seems to obscure the underlying mechanism of verse generation. In order to reveal such a dynamic dimension of verse making, one may restructure the inventory in terms of derivational processing, as follows: Table 1.1. Formation of basic verse types Initial position

Second position

Third and final positions

Whole configuration

By promotion

Verse class

Verse type

/  /  /  ×  × 

×  ×  /  /  / 

/× ×/ ×× ×/ /×

/ × / × */ × × / / / × × × / × / × / / ×

—  /\×/ —  —  — 

A  E  D  B  C 

A1 E  D1 B1 C 

Taking the first row as a representative example, the above table reads: put / in the first slot, × in the second, and / × in the last two, to obtain the configuration / × / ×, verse class A, prototypically realized as verse type A1 (see below). The second case needs an additional process of promoting the first drop × (second position) to a heavy one \, so that a well-formed configuration / \ × / may substitute for the otherwise ill-formed sequence */ × ×  /. The extra implementation of metrical reinforcement means that the whole process of derivation is more complex than those responsible for the other four types. This derivational complexity provides a formal account of why type E among others occurs with minimal incidence. The derivational representation also brings to light a major dichotomy of the basic verse types to two groups according to the identity of the verse-initial position, lift-initial (A, D, E) or drop-initial (B, C). As will be substantiated in subsequent chapters, this major division of verse types predicated on verse-initiality (verse opening) has profound significance in the metrical organization. While the metrical strengthening of a drop to a heavy one is uniquely involved in deriving one of the five basic types (type E), comparable operations of drop promotion are found responsible for constructing heavier variants out of some of the basic configurations. Operating on each drop in type A (/ × / ×, hereafter referred to as type A1— the basic variant of verse class A—in distinction from its derived variants) and type D (/ / × ×, hereafter referred to as type D1, in distinction from its derived ones), a metrical strengthening gives rise to the following increased types: (2)    

Increased verse types Type A2a / \ / ×; type A2b / × / \; type A2ab / \ / \ Type D2a / / \ ×; type D2b / / × \

8 

 1 Introduction

Thus, an increased type is derived from its basic counterpart through substitution of a heavy drop for a normal one. As should be clear, the classifier “2” designates an increased type, that is, a derived variant resulting from metrical strengthening of a drop in general; the labels “a” and “b” indicate that the promotion at issue involves the first and the second drop, respectively. Correspondingly, the basic types that provide a basis for metrical strengthening are distinguished by the label “1,” namely, types A1 and D1, noted above. A secondary process of metrical adjustment that is implemented on the basic types may proceed in the opposite direction, that is, by reducing a lift to a drop rather than strengthening a normal drop to a heavy one. Specifically, the first lift may be replaced by a (normal) drop, producing types A3 (× × / ×) and B3 (× × × /). The designator “3” corresponds to a reduced type that is constructed by such metrical demotion. Thus, the whole system of verse types may be organized with maximal elaboration, as follows: Table 1.2. The system of verse types Class

Basic type

Increased type

Reduced type

A  B  C  D  E 

A1 B1 C  D1 E 

A2a, A2b, A2ab —  —  D2a, D2b — 

A3 B3 —  —  — 

Inasmuch as both the increased and reduced types are derived through an extra mechanism of transforming the default composition comprising the maximally differentiated positions, two lifts and two normal drops, they must be counted as marked configurations. These marked verse types—increased or reduced—are distributed differentially between the a-verse and the b-verse, as most eloquently testified in Beowulf (Suzuki 1996: 65–68, 95–110), and somewhat less conspicuously in the Heliand (Suzuki 2004: 66–68, 125–136). Regardless of the extent of differentiation, it holds generally true that the a-verse is more accommodating: the marked types are more likely to occur in it than in the b-verse on statistical grounds. Moreover, which particular derived types are in function is an empirical issue that requires substantiation for each work of poetry (cf. Suzuki 1996: 375; 2004: 177; 2014a: 203). Accordingly, the major issues addressed in the synchronic parts of the subsequent investigations into the meters of Genesis A and Genesis B will be to determine in rigorous terms the identities of verse types, their varying linguistic realizations, and their underlying organization into the paradigmatic system. Drawing on the analytical framework outlined above, I will provide a typological survey of the organization of Old Germanic meters in their structural foundations. Of

1.3 Variation on the principle of four metrical positions in Old Germanic meter 

 9

central concern will be the two major metrical traditions: Beowulf (Beo; Old English; Fulk/Bjork/Niles 2008) and the Heliand (Hel; Old Saxon; Behaghel/Taeger 1996). As stated in section 1.1, the whole metrical organization of the Old English Genesis and its constituent meters of Genesis A and Genesis B would have been constructed in the intersection of the Old English and Old Saxon verse traditions. It should be noted in parentheses that the notion of typology just mentioned concerns not so much the one of meters from a universal perspective (e.g., Lotz 1960; Wimsatt 1972; Fabb/Halle 2008; Aroui/Arleo 2009) as the localized one in a strictly Old Germanic context. In this sense, the notion corresponds to the archaeologists’ use rather than the linguists’: for example, the typology of Anglo-Saxon button brooches (Suzuki 2008).

1.3 Variation on the principle of four metrical positions in Old Germanic meter: The verse size While the principle of four metrical positions—itself being stochastic rather than categorical in nature—largely serves as a foundation for verse composition in Old Germanic, the exact status of the principle varies from meter to meter in its binding force, namely, the extent to which it tolerates deviations from the prescription it lays out. There are two kinds of verses inherent in Old Germanic meter that may potentially count as larger in size than is prescribed by the principle of four metrical positions per verse: verses extended with anacrusis on the one hand (e.g., Beo 758a; see [3] below), and class D* (/ × / × ×)—or its variants, types D*1 (/ × / × ×), D*2a (/ × / \ ×), and D*2b (/ × / × \) in Beowulf (Suzuki 1996: 103–112; e.g., Beo 1749a; see [3] below)—on the other, which is expanded with an additional drop standing between the otherwise consecutive lifts in class D (/ / × ×)—or its variants, types D1 (/ / × ×), D2a (/ / \ ×), and D2b (/ / × \) in Beowulf (Suzuki 1996: 95–103). In addition, these two expansions can be conjoined, resulting in still longer verses that could be regarded as consisting of six positions (e.g., Beo 2936a; see [3] below): (3)    

Beo 758a Gemunde þā se gōda [xPxxxPx]4 (type A1 with anacrusis) Beo 1749a gȳtsað gromhȳdig [PxPSx] (type D*2a) Beo 2936a Besæt ðā sinherge [xPxPSx] (type D*2a with anacrusis)

4 P = primary-stressed long syllable; p = primary-stressed short syllable; S = secondary-stressed long syllable; s = secondary-stressed short syllable; x = unstressed syllable, long or short; # = word boundary (specifically marked in limited contexts).

10 

 1 Introduction

In the face of this range of variability, one has three parameters at one’s disposal by means of which to differentiate the varying strictness of the principle of four metrical positions: (i) the incidence of anacrusis; (ii) the occurrence of class D* relative to class D; (iii) a combination of (i) and (ii), namely, the expansion of class D* with anacrusis. Given that anacrusis applies to classes A (excluding type A3), D, and D* commonly in Beowulf and the Heliand, one may compare the distribution of corresponding configurations with or without anacrusis in the two poems (Table 1.3). However, since the realization of anacrusis by prefixes does not constitute an autonomous metrical position in Beowulf (Suzuki 1996: 337–340), prefixes cannot serve as a proper indication of the extent to which the principle of four positions per verse tolerates anacrustic composition of overlong verses in terms of the number of constituent verses. Accordingly, anacrusis as realized by nonprefixal material, that is, independent words—as exemplified in (4) below—is the key to understanding these apparently overlong verses. The relevant figures are indicated under the heading “nonprefixal” in Table 1.3. Table 1.3. Occurrences of anacrustic and nonanacrustic configurations (classes A, D, and D*, excluding type A3)  

Anacrustic

Nonanacrustic

Total

Beowulf Beowulf (nonprefixal) Heliand Heliand (nonprefixal)

 86 (2.68 %)  19 (0.60 %) 1386 (26.41 %) 1027 (21.01 %)

3125 (97.32 %) 3125 (99.40 %) 3861 (73.59 %) 3861 (78.99 %)

3211 (100 %) 3144 (100 %) 5247 (100 %) 4888 (100 %)

In any event, anacrusis is far more likely to occur in the Heliand than in Beowulf (p