Overwriting, deletion and erasure: exploring the changes in the Ormulum manuscript.


Nils-Lennart Johannesson



The project


Over the last five years or so, I have become increasingly involved with a text known as the Ormulum, a twelfth-century collection of homilies (or sermons). The text is preserved in a single manuscript (MS. Junius 1) in the Bodleian Library, Oxford. The only modern edition of the text appeared in 1878, an edition which is far too unreliable to be used for any investigation of the language in the manuscript. So for a study of variation in the use of the third person plural pronouns in the Ormulum (such as heore (of native English origin) vs. þe33re (borrowed from Old Norse), both meaning ‘their’) I bought a microfilm of the manuscript from the Bodleian Library. To be able to use the material for other purposes as well without having to go through the text on the microfilm again, I decided to convert the text of the manuscript to a database. That turned out to be a task that has kept me off the streets for the last four years, even though I could make use of various triumphs of modern technology, such as microfilms, photocopiers, scanners and laptop computers, none of which were available to earlier scholars working with this text.


By the summer of 1996 I had finally produced a database which contained all the text I could read off the microfilm. However, due to the many changes in the manuscript (including deletion and erasure of text), the database also contained a very great number of question marks where it was impossible to tell from the microfilm what the text had originally been and/or had been changed to. The obvious next step was then to check my transcript (the text in my database) with the actual manuscript. Thanks to the cooperativeness of the English Department, the Arts Faculty of NTNU, and the Bodleian Library, I have had the opportunity to work full-time on this project in Oxford this semester. Many thanks to all who made this possible; what follows is a brief account of what I have been occupied with during my stay in Oxford.



The library


Duke Humfrey’s Library is the oldest part of the Bodleian Library. It originated as a donation of a collection of manuscripts to Oxford University by Humfrey, Duke of Gloucester (younger brother of Henry V) in 1444. The building which was to house this collection, a room added on top of the Old Divinity School, was completed in 1488. Today, Duke Humfrey’s Library houses the Bodleian’s collection of Western manuscripts and books printed before 1660. All those old leather-bound volumes, kept in the original fifteenth-century book presses, create a very special atmosphere. The ceiling is covered with panels bearing Oxford University’s coat of arms, and the walls above the book presses are hung with 16th century portraits of knights, noblemen and clerics (and a few ladies). Two busts of Sir Thomas Bodley (from 1605) and Charles I (from 1636) complete the decorations. The modern system of low-voltage lamps lighting the ceiling is a wonder in itself — you don’t see the lamps or the light beams from the lamps, you just see the beautifully decorated ceiling hovering in the brilliant light high above you.[1]


It is in this kind of environment that I have had the privilege to investigate the manuscript. The library is, naturally, concerned about minimising wear and tear of the manuscript, but I was granted access to it for four weeks, which I have spread over the semester.[2] Due to the complex nature of the manuscript, I have certainly needed every minute of that time. For every week spent with the manuscript, I have also needed about four weeks at my computer systematising all the information I have noted down in pencil in my transcript of the text. After my time here in Oxford, I estimate that the analysis of the metre and grammar, as well as the preparation of a new edition of the text, will keep me occupied for the next five years.



The text


The Ormulum is a collection of homilies arranged as a life of Christ, written in the East Midland dialect of Middle English after the middle of the twelfth century. The author identifies himself as an Augustinian canon called Orm, in all likelihood working at the abbey in Bourne in southern Lincolnshire,[3] some ten miles north of Peterborough. The homilies, as well as the Dedication to Orm’s brother Walter, who commissioned the work, are written in verse in a metre known as the septenarius, each verse having fifteen syllables divided into seven feet.[4] By literary critics Orm has been severely criticised for the monotony of his meter (not true, actually) and the dreariness of his exposition (rather more justified). Language historians, by contrast, have embraced him for his ingenious spelling system (which allowed him to mark vowel quantity in closed syllables), which he applied with a degree of consistency that is almost without parallel in the Middle English period. For that reason, the Ormulum is our best source of information about vowel quantity in early Middle English.


The Ormulum is a huge work in more than one sense. The manuscript measures 21 by 51 centimetres (although the pages vary a good deal in size and shape; the quality of the parchment, by contrast, is invariably poor). The surviving text runs to over 20,000 lines in the 1878 edition, and this is only a fragment of the original work: there are several pages missing from the surviving part of the manuscript, and that part breaks off in the middle of homily 32, whereas the table of contents (itself incomplete) lists the texts of 242 homilies. After entry 50 in the table of contents, a collaborator of Orm’s (commonly referred to as Hand C) has written Huc usque .i. uol. (‘Here ends volume 1’); it seems reasonable, then, to assume that the whole work was planned as five volumes with a total of 250 homilies. In the Dedication, Orm tells his brother Walter (who was not only his brother in the flesh, but also a brother in the same monastery) that they should give thanks to God since the whole project is now brought to an end, so we have to allow for the possibility that he actually wrote all the 250 homilies. My own experiments with copying parts of the text with Orm’s script suggest that the copying of the surviving fragment must have taken at least six months, or 45 years for the five volumes. It is more difficult to estimate how long it took him to compose all that text, but a couple of decades does not seem unreasonable.


Orm makes no direct references to the contemporary situation in England in his homilies. However, one passage which may have been prompted by recent events can be found in Orm’s exposition of John 4.13, where he says


(1)               & ec þe laferrd crist attflæh;

                      Forr þe to gifenn bisne.

                   Þatt tu mihht flen. & berr6enn swa.

                      Þi lif wiþþ godess lefe;

                   & shunenn þa þatt wilenn þe.

                      Wiþþ utenn gillte cwellenn;

                   5iff þu ne mihht nohht habbenn 3ët.

                      God lusst. god mahht. god wille.

                   To þolenn marrtirdom forr crist;

                      & forr þe rihhte læfe.


(‘And the Lord Christ also fled [from the Pharisees] to make it clear to you that you may flee and thus save your life with God’s permission, and keep away from all those who want to kill you without guilt, if you do not yet wish to suffer martyrdom for Christ, and for the true faith.’)


The prospect of suffering ‘marrtirdom forr crist’ in the East Midlands was a very real one in the 1140’s: Geoffrey de Montfort, earl of Essex, seized Ramsey Abbey (13 miles south-east of Peterborough) and expelled the monks in 1143; from this stronghold in the Fen Country he terrorised the countryside for over a year, until he died excommunicated in 1144.


In its long account of the atrocities committed during the Anarchy, the civil war during Stephen’s reign 11371154, the Peterborough Chronicle says about robber barons like Geoffrey de Montfort that “they extracted payment from the villages and called it ‘protection money’. When the poor people had no more to give, they plundered and burned all the villages ... they spared neither church nor churchyard, but took all the goods they found there, and then burned the church ... When two or three men came riding to a village, the villagers fled from them, believing they were robbers ... People said openly that Christ and his saints slept.” Against such a background, it is perhaps only natural that Orm, who presumably started writing in the late 1140’s or early 1150’s, abandons his normal technique of fairly literal translation of the Gospel text in his rendering of John 3.14. When the Roman soldiers ask St. John the Baptist what they should do to be saved, Orm has the Baptist launch into a long impassioned speech (using both internal rhyme and alliteration), listing the various kinds of atrocities they should abstain from.


One of the striking features of the Ormulum is the highly idiosyncratic spelling system that Orm devised and applied with remarkable consistency throughout his long text. Its chief characteristic is the doubling of a consonant after a short vowel in a closed syllable, as in affterr ‘after’, comm ‘came’, þurrh ‘through’. The motivation behind a very great number of changes in the manuscript was clearly a wish to eliminate variant forms[5] in an attempt to make the morphology and syntax as uniform as the orthography.


Much could be said about the content, metre, vocabulary and grammar of the Ormulum, but since my work this semester has been confined to the ink-and-parchment level, I will conclude with an account of the way Orm went about making his changes and give a hint of what one may find if one takes the trouble to look for his original forms.



Changes in the manuscript


The manuscript can be seen as one huge illustration of writing as process rather than product.[6] Orm obviously started out copying his text from an earlier draft, making mistakes as he went along (e.g. skipping words or parts of words). In some cases he spotted his mistake at once and corrected it then and there; other mistakes were discovered and corrected later, and a few escaped notice altogether. When Orm had finished the text, he then started revising it: passages were removed, new passages were added, either in the margins or on additional parchment sheets sewn into the manuscript. He then went over the manuscript a number of times, editing various details of spelling and grammar. It is possible to trace this development, because his handwriting deteriorates gradually, as his eyes grew dimmer and fingers stiffer with increasing age. Towards the end of the editing phase he could only just produce recognisable letters.


One may also note that Orm, who had been consistently progressive in his choice of variant forms during the editing stage (i.e. switching to forms which have more in common with Modern English than with Old English, e.g. deofel > defel ‘devil’, icc habbe > icc hafe ‘I have’, heore > þe33re ‘their’), becomes reactionary towards the end: at the stage when he produces barely recognisable letters, wimmann ‘woman’ is changed to wifmann (cf. OE wifman), and the forms a ‘a’ (indef. art.), na ‘no’, mi ‘my’ and þi ‘thy’ are changed to an, nan, min and þin, recreating the Old English uninflected forms.


The means available to Orm for making changes in the text he had written were overwriting, deletion and erasure.


Overwriting means quite simply that you write a new character over an existing one. For example, to change an infinitive form ending in -enn to a past participle ending in -edd, you simply write the ds on the ns, and leave it at that. This kind of change is usually easy to spot, since in most cases it results in oddly shaped letters, or at any rate shapes which Orm never uses otherwise. In contemporary manuscripts, two variants of the letter ‹d› are used: one which looks like the modern printed ‹d›, and one which looks like the figure 6 turned back to front. Whenever Orm uses the first type, its a certain sign that he has got something to hide.


Deletion is the technique of drawing an ink stroke through the offending text passage. In many cases, enough of the text is showing above and below this stroke to make it possible to read the deleted text without difficulty. In other cases, however, Orm was not content with just indicating to a possible future copyist that a passage was to be omitted, but seems to have made a special effort to make his original text illegible. He then either scraped off whatever could be seen outside the deletion stroke, or covered the text completely with ink. In the latter case, the outline of the ink blot will typically give a good clue to the type of letters (ascenders[7], descenders, or neither?) deleted. And in most cases, patient scrutiny of the passage with light coming in from various angles will pay off either the letters can be seen through the ink, or the ink of the original letters left so much pigment on the parchment that the letters can be seen rising slightly above the surface of the page, even under the cover of the deletion stroke. Very few cases of deleted text defy identification altogether.


Erasure, finally, involves the use of a knife to scrape off the offending text from the parchment. This scraping can be done in two different ways: superficial erasure just removes the pigment[8], leaving pink marks on the parchment which can easily be read (unless they have been covered by too much new text), while deep erasure removes the surface layer of the parchment as well, thereby removing all traces of the ink. However, all may not be lost even in the latter case. Deep erasure results in a very rough surface, not well suited for writing on, so in many cases Orm tried to avoid this by using the tip of the knife to remove a letter without damaging the surrounding parchment surface. This means, of course, that the outline of the letter can still be seen, even if new text has been written over the erased passage. Nevertheless, a number of erased text passages have been lost for good. Trying to make sense of faint traces of erased text, overwritten in black ink, was easily the most time-consuming and tiring part of my work with the manuscript. However, it did pay off in a number of cases.


Deletion and erasure were used for partly different purposes. Deletion, which can vary in scope from a single letter to nearly a whole page, was used to make changes in the later stages of the revision process. These changes range from the trivial (such as the substitution of Þatt witt tu wel to soþe You should know that for a truth 40 times for Swa summ þe goddspell kiþeþþ As the gospel says or Swa summ þe boc uss kiþeþþ As the book tells us, when the passage he is referring to does not actually occur in the Gospel) to the more interesting cases where Orm apparently was dissatisfied with the content of the deleted passage. One example of this type is the much-debated description of the altar in the Temple in Jerusalem, where Orm has deleted several passages with extra heavy ink strokes. Most of the deleted text was identified by R. Burchfield in the 1950s;[9] I have now been able to fill in the missing parts. The reason for these changes was obviously a wish to avoid any reference to hali3 bræd (sacred bread) or hall6edd bræd (consecrated bread), possibly because he feared that this might mislead his congregation into believing that the Jews celebrated Communion.


Erasure was primarily the technique for the immediate correction of errors. Often Orm realised after writing a couple of letters that he had skipped a word, or skipped a line or verse, as he was copying from his rough draft. He then put down his quill, picked up the knife, erased as much as he needed, and wrote the correct text. These cases are easily recognisable because the current text runs from the erasure onto untouched parchment without a break and without being crowded in or stretched out to fill a gap. In other cases he discovered the omission of a word or syllable only after completing a page. He then erased the whole line and rewrote the line with the missing bit added. This gives the line a characteristic crowded appearance, so even when the original text is illegible is it possible to see what has been going on.


But erasure could also be used when Orm wanted to make changes in the text, either in order to modify his message or in order to make grammatical modifications. An example of the former type is when he first described Job’s illness as micclelikess adle ‘the affliction of dropsy’ and then erased micclelikess and wrote an full atell ‘a very foul (affliction)’ instead. The latter type can be found on a page where he made extensive changes to ensure that the verb don ‘to make (somebody do something)’ is followed by an infinitive preceded by to (since he had to make sure he had the right number of syllables per verse, he couldn’t just add the to without any further changes). Thus, discoursing on humility, he first wrote the text as it stands in (2a) below; at a later time, he erased the verb stem and introduced the stem of a synonymous verb which was one syllable shorter (to make room for to), as shown in (2b).


 (2) a.       Þiss hæfedd mahhte doþ þe wel.               b.

                      5iff itt iss i þin herrte.

                   Forrwerrpenn derewurrþe shrud;                 To shunenn derewurrþe shrud;

                      & derewurrþe mæless.


(‘This cardinal virtue makes you, if it is in your heart, reject costly dress and costly meals.’)


The most important lesson to be learnt from these changes is that we have to understand what Orm is doing in his manuscript before we can start drawing conclusions about the variety of English represented there, as my final example will illustrate.


In his discussion of the early history of h-dropping (as in the use of ’im for him) in the Cambridge History of the English Language, James Milroy points out that h-dropping in the early Middle English period was most common in texts from the East Midlands.[10] The Ormulum has always stood out as an exception: Orm apparently had complete control over his h’s, despite the fact that he came from the heart of the h-dropping area. It turns out, however, that he was just better than other scribes at covering his tracks. His dialect (southern Lincolnshire at the middle of the twelfth century) was obviously characterised by h-dropping: among the originally written forms, later erased (or otherwise modified), we find imm (for himm < OE him), iss (for hiss ‘his’ < OE his), affde (for haffde ‘had’ < OE hæfde), willke (for whillke ‘which’ < OE hwylc) and allflinngess (for hallflinngess ‘partly’ < OE (Anglian) half + OE -ling). The last word is particularly interesting since it occurs at the beginning of a verse and is therefore written with a capital A, so Orm must in all likelihood have been aware of what he was doing. In the end, however, his urge for uniformity got the better of him, and he erased the word and rewrote it with a capital H.


This is just one small example of what can emerge if one has a chance to observe somebody’s language habits in unguarded moments. The Ormulum is a fascinating document in that it is possible to get glimpses of the amount of variation that must have been possible in Orm’s dialect, although he himself aimed at a consistent and invariant use of the English language of his time and region.


So what can we learn from a study of these changes?


The most widely discussed (and, in my view, most misunderstood) case involves words which in Old English had the diphthong /eo/ (long or short)


The Old English diphthong /eo/ (long or short), represented by the digraph eo, became monophthongized to /O/ and further unrounded to /e/ in early Middle English. Orm, who otherwise is extremely consistent in his spelling, uses parallel forms with eo and with e in words with this vowel (e.g. beon/ben ‘be’, deofel/defel ‘devil’, heofen/hefen etc.). After about 14,000 verses Orms gives up the forms with eo, and, furthermore, goes over the manuscript with the knife and erases all the o’s in the eo words that he finds. Consequently, it has been generally assumed that wherever the letter e is followed by an erased letter, Orm’s original text had eo. This leads to problems in



[1] Sadly, the roof-timbers have been found to be infested with death-watch beetle. The University is apparently having problems financing the necessary repairs: the closure of the library, which was earlier announced for the autumn of 1997, has recently been postponed for a year.

[2] I am grateful to Dr. B.C. Barker-Benfield, Senior Assistant Librarian, for granting me permission to study the manuscript.

[3] For a discussion of the reasons for this localisation see M.B. Parkes, ‘On the Presumed Date and Possible Origin of the Ormulum: Oxford, Bodleian Library, MS Junius 1’ in E.G. Stanley and D. Gray (eds.), Five Hundred Years of Words and Sounds: A Festschrift for E.J. Dobson (Cambridge, 1983), pp. 115127.

[4] One consequence of Orm’s strict adherence to the requirement for 15 syllables per verse is that whenever he deletes a syllable or two, he has to add something to compensate for the loss of syllables. He normally relies on the addition of monosyllabic adverbs such as a33 ‘always’, þa ‘then’, þær ‘there’, or he exploits variant forms of words (e.g. ‘if’ could be either 3iff or 3iff þatt).

[5] For details, see R.W. Burchfield, ‘The Language and Orthography of the Ormulum MS.’, Transactions of the Philological Society 1956, pp. 56–87.

[6] Most of folio 85 r. of the Ormulum can be studied on the cover of the 1996 annual report of the English Department, NTNU. This extract gives a somewhat misleading impression of the manuscript, since the text here is unusually ‘clean’. In the bottom left corner, however, it is possible to see the word a33 ‘always’ added in a balloon to compensate for the loss of a syllable when gode ‘good’ was changed to god.

[7] A brief explanation of terminology: the vertical strokes in letters such as n, m, and u are called minims. Letters such as a, c, e, m, n, o extend vertically between the baseline and the headline. Letters such as b, d, and h, extending above the headline, have ascenders, and letters such as p and q, extending below the baseline, have descenders.

[8]  The ink used in the writing of the manuscript consisted of pigment (typically lamp soot) suspended in a liquid which served to etch the surface of the parchment and bind the pigment to that surface. In places where the pigment has been lost (either due to wear and tear near the edges of a sheet, or due to poorly made ink and/or poor quality of parchment), it is still possible to read the text from the etch marks, standing out pale pink against the off-white parchment.

[9] Burchfield, op.cit.

[10] J. Milroy, ‘Middle English Dialectology’, p. 199, in Cambridge History of the English Language, Volume II 10661476, ed. by N. Blake (Cambridge, 1992), pp. 156206.