The Scansion of Classical Old English Verse

Acknowledgements

The publication of this book would not have been possible without the generosity of the New College Ludwig Fund for the Humanities, Oxford, and the support of the Ramón y Cajal programme of the State Research Agency of Spain (Grant ID: RYC2022-035374-I).

I began this project in 2018 as part of my research for CLASP: A Consolidated Library of Anglo-Saxon Poetry. The online Excel database presented here is an expansion of the one developed by Andy Orchard, the Rawlinson and Bosworth Professor of Anglo-Saxon at Oxford, prior to my joining CLASP. When I began my work as a researcher there, Professor Orchard kindly shared his original database with me. Derived from raw texts in the Oxford Text Archive and generated by Professor Orchard using the Oxford Concordance Program (OCP), the database contained the text of The Anglo-Saxon Poetic Records, organized into individual Excel forms by verse, along with an automatically generated scansion. This was incredibly helpful, as it allowed for more efficient handling and analysis of the data, saving me a significant amount of time. I am deeply grateful to Professor Orchard for permitting me to use this database, as well as for his ongoing encouragement and support.

I would also like to express my thanks to Robert D. Fulk, Class of 1964 Chancellor’s Professor Emeritus of English at Indiana University, and to Nelson Goering, Postdoctoral Fellow in Medieval Studies at the University of Oslo. I have greatly benefited from their expertise in metre on numerous occasions.

Trevor Powell, Emeritus Fellow at New College, ensured I never grew complacent during our college lunches, with his persistent (and occasionally pointed) questions about the project. Though sometimes probing, these questions ultimately helped me stay focused and on track.

Finally, I owe a very special thanks to Mark Griffith, Richard Ellmann Fellow and Tutor in English at New College, for his extraordinary patience in always providing me with precise and insightful answers to the countless queries that arose throughout this project. His immense knowledge of all things related to Old English, combined with the seriousness and dedication with which he approaches his work, serve as an example I deeply admire and aspire to emulate. Without his encouragement and friendship, this work might never have reached its conclusion.

Introduction

This volume consists primarily of an index of scansion, a set of metrical tables, and an online metrical database drawn from what is here defined as ‘classical’ Old English verse – that is, most of the poetry found in the first four volumes of The Anglo-Saxon Poetic Records, composed between the 8th and 10th centuries according to the principles of Sieversian metrics. Poems such as Genesis B (a translation from Old Saxon) and Maxims I (a gnomic composition with idiosyncratic metre) follow metrical rules that are related to, but significantly different from, those governing works like Beowulf or the Cynewulfian canon. Nevertheless, they have been included for comparative purposes and to preserve the integrity of the ASPR. The metrical database itself – upon which both the index of scansion and the statistical tables presented in this volume are constructed – is available exclusively online in sortable Excel format, as a digital supplement to the physical book. The volume as a whole is intended as a reference tool for scholars of Old English, both for non-specialists in metrics – who may find the index particularly useful – and for specialists, for whom the tables and the online database are expected to be of value.

The total number of half-lines or verses analysed in this volume is 40,372, which represents approximately 67% of the surviving corpus (60,000 verses). The book’s most obvious predecessor, B. R. Hutcheson’s 1995 monograph, Old English Poetic Metre, was based on the analysis of 26,000 verses, or roughly 43% of the corpus. Hutcheson, moreover, employed a unique system of scansion, distinct from the widely recognized labels and terminology developed by Sievers, and, more crucially, the database underlying his analysis was never made available by the publisher – likely due to the limitations of digital resources in the 1990s. The absence of this database has been a notable limitation, as highlighted by Mark Griffith in his review of Hutcheson’s work:

The blurb at the beginning claims that the database is also available from Boydell & Brewer, but in fact the publishers have as yet no date in mind for its publication. Accordingly, I have been unable to check the statistics that underpin Hutcheson’s methodology. Those readers who like to judge the information for themselves will find, likewise, that the book is not as useful as it might otherwise have been.

The primary objective of this volume is to supersede Hutcheson’s work as a reference resource, incorporating a Sieversian analysis of a significantly larger number of verses, along with statistical tables for each individual work and, importantly, the database upon which these materials are based. This will allow readers to verify the underlying data for themselves. It is essential to note, however, that I hold a deep respect for Hutcheson, to whose memory this book is dedicated. His untimely death left the field of Old English philology deprived of many valuable contributions, and I am confident that, despite the publication of this book, his work will remain an important resource for the study of Old English metrics and textual criticism – especially for the discussion of particularly difficult verses.

A note on how to use the materials here presented is now in order. I should emphasize two points regarding my scansion: first, I employ the labels of traditional Sieversian metrics (see, for example, the section on versification in Pope-Fulk 2001: 144–48); second, I have aimed for a generally conservative approach. Take, for example, Beowulf 25a, in mægþa gehwære. This verse is irregular, comprising five metrical positions – in is not typically anacrustic in that poem, and if it were, the verse should then exhibit double alliteration. I believe that the poet originally composed a regular four-position Type B verse, in mægþa gehwæm, and that a later scribe disrupted the metre by substituting the analogical feminine pronoun gehwære. Nevertheless, I have chosen to follow the manuscript reading and classify it as a remainder (‘rem’.). This decision reflects the book’s purpose: not to offer a new edition of classical Old English verse, but to serve as a metrical companion to the ASPR. Readers should note, however, that several such verses (especially among the remainders) have plausible editorial emendations proposed in the literature.

As noted above, my analysis is grounded in the principles of Sieversian metre. In the ‘Scansion’ column of the online database, I use traditional Sieversian symbols: ‘/’ for primary stress, ‘\’ for secondary stress, and ‘x’ for unstressed positions. Resolution and its suspension are marked by ‘~+’ and ‘~’, respectively. I have also incorporated several elements from Blissian metrics, which I believe offer valuable insights not fully accounted for in Sievers’s system. One such addition is the treatment of the caesura. Sievers regarded the caesura as a theoretical construct determined solely by the metrical type of a verse. In contrast, Bliss treated the caesura as syntactically determined. Consider, for example, the half-lines husa selest and geong in geardum. Both are Sieversian Type A. Accordingly, Sievers would place the caesura in both instances after the first unstressed syllable – that is, husa | selest, and geong in | geardum – thereby dividing each half-line into two trochees, which is at odds with syntactic structure. Bliss, by contrast, locates the caesura in accordance with syntactic boundaries: husa | selest (caesura in position 2) versus geong | in geardum (caesura in position 1). Notably, the Blissian caesura tends to correlate with alliterative patterns – double alliteration being nearly obligatory when the caesura is in position 1, and optional in position 2 – further justifying its inclusion in this analysis. (For further discussion, see Bliss 1967: 36–39.) Additionally, the database classifies verses according to Bliss’s categories of ‘heavy’ – those with three stress-words rather than two (e.g. wæpen wundrum heard, a Type D verse) – and ‘light’, with only a single stress-word (e.g. þone leofestan, a Type C with alliteration on l).

The other principal parameters – or columns – in the database are double alliteration (‘DA’), anacrusis, and whole-verse compound (‘WVC’), the last of which refers to verses composed entirely of a single stress-word (e.g. hilderinca). These categories should be transparent to most readers. There is also a ‘Notes’ column, which indicates any slight deviations from the manuscript reading made during scansion. For example, Genesis A 17a, demdon drihtenes, is here analysed not as a Type D, but as a Type A with syncope of medial -e-. Similarly, Genesis A 2a, wereda wuldorcining, is classified as a Type D*2, with wuldor scanned as a monosyllable (from an earlier form *wuldr). These cases are noted in the ‘Notes’ column as ‘Syncope’ and ‘Non-parasiting’, respectively. In contrast, in Christ and Satan 6a, wæter and wolcn, the word wolcn is scanned as a disyllable, and the corresponding annotation is ‘Parasiting’.

I have also used the ‘Notes’ column to subclassify those verses labelled as ‘remainders’ (‘rem’.). These annotations help to clarify the specific metrical irregularities that led to remaindering. The subcategories that I have used include:

• A3(b) in off-verse
• Alliteration
• Anacrusis and caesura
• Anacrusis with single alliteration
• B3 in off-verse
• Caesura
• Catalectic
• Double alliteration in off-verse
• D* in off-verse
• D* with single alliteration
• Extra syllable
• Five positions
• Heavy drop
• Isolated hypermetric
• K1
• Name
• Suspension of resolution
• Unstressed infinitive
• Other

Many of these will be readily understood, but others require further comment. ‘Alliteration’ is used for alliterative problems of different sorts. ‘Anacrusis and caesura’ refers to verses like on hwilce healfre, where both breath-groups are the same size and anacrusis is therefore not expected (see Bliss); such verses are generally not considered hypermetric, as the excess amounts to a single syllable. The label ‘Caesura’ marks problems of caesural placement, as in frea on forþwegas, where a Type D* verse should show the caesura after the first unstressed syllable, not before it. ‘Catalectic’ applies to unambiguous three-position verses such as on genimeþ, in contrast to cases like wæter and wolcn, discussed above, for which a regular four-position analysis is still viable. ‘Extra syllable’ identifies verses like mætra on modgeþanc, since one would expect the extra drop of Type D* to be occupied by just one syllable (though elision may be a plausible alternative). Verses such as se yldesta wæs are marked as ‘Heavy drop’ because Type E normally does not accommodate anacrusis. A verse is considered an ‘Isolated hypermetric’ not only when it occurs without a corresponding hypermetric half in the same line, but also when no other hypermetric verse appears in the immediately preceding or following line. ‘K1’ flags violations of Kuhn’s First Law, as in Christ and Satan 429a, wolde helwarum, where the finite verb wolde is unstressed but displaced from its normative position at the beginning of the verse-clause. ‘Suspension of resolution’ is used for verses like gestaþelode, which show problematic short lifts (like -lo-) immediately following resolved lifts (-staþe-). Finally, I have remaindered most verses containing non-Germanic names (marked ‘Name’), even when their scansion is more or less regular. The label ‘Other’ is used for verses that exhibit anomalous features not adequately captured by the other subcategories.

So much for the digital database, which was the first of the components to be produced, and also the most foundational, as it underpins the two principal elements of the physical book: the index of scansion and the set of statistical tables. The index has been modelled on that found in Bliss’s monograph (1967: 139–61), with one key modification: since Sieversian labels (unlike Bliss’s) do not indicate the position of the caesura, this has been recorded in a separate column where relevant. (In Types C and D, for example, the caesura is always in the same position and therefore not marked.) The statistical tables present the incidence of metrical types in the on- and off-verses of each poem over fifty lines. There are individual tables for each riddle exceeding fifty lines, as well as a composite table for all of them treated as a single poem. Figures are presented as raw numbers, except that percentages – highlighted in bold and accompanied by their corresponding raw values in parentheses – are given for the basic types (A1, B1, C1, D1, and E), and for metrical subtypes where the percentage is five or more. Readers should bear in mind that these statistics reflect my own analysis, which necessarily involved a number of interpretive decisions. For example, I have scanned finite verbs outside Beowulf in the same manner that Bliss does in his Beowulf monograph, and – unlike Bliss – I have treated so-called tertiary stress as metrically significant. Other metrists might have adopted different approaches to these issues; nevertheless, such decisions were unavoidable. Even so, the tables should offer a reliable indication of general metrical and stylistic trends across the poems – particularly as, to my knowledge, no comparable resource has previously been compiled.

The physical book concludes with three final lists: the first records all verses interpreted as hypermetric (excluding isolated instances, as previously defined); the second list includes verses with anacrusis in the on-verse, while the third does the same for the off-verse. In general, instances of anacrusis in these lists consist of prefixes or monosyllabic elements. A total of 636 verses – approximately 1.6% of my corpus – have been identified as exhibiting anacrusis. Of these, 548 appear in the on-verse (constituting 3% of on-verses), and 88 in the off-verse (0.4%). The number of verses analysed as hypermetric stands at 831, or roughly 2% of my corpus, with 405 occurring in on-verses and 426 in off-verses.

I would be grateful to hear from readers who notice any inconsistencies or errors, or who have suggestions for improvement, as it is my intention to keep the database up to date.

Rafael J. Pascual
Granada, June 2025