Personal names in medieval libri vitæ as a sociolinguistic resource

2.1 The liber vitæ of Reichenau

The liber vitæ of Reichenau is the most comprehensive of all the libri vitæ and contains more than 38,000 personal name entries from the ninth to the fourteenth century. It was composed as a liber vitæ between 823 and 825 (cf. Zettler 2010: 65) including name lists that were collected following the synod in Attigny in 762, where leading clerics from various Continental monasteries agreed to memorise each others’ deceased. For this oldest layer that is documented on folios 15v–58v, a table of contents^[5] was created at the beginning of the codex. Each list is then introduced with the order number from the capitula and a title naming the respective monasteries in red ink, as shown in Figure 1. The lists are all neatly arranged in columns. In addition to these clerical lists, the oldest layer also contains lists with secular rulers, such as emperors and kings, and other benefactors and friends of the monastery on folios 62v and 63r (Autenrieth et al. 1979: XXVII). In later years, the systematic structure of the oldest layer was no longer followed and names, both of clerics and secular people such as friends, sponsors of the monastery and groups of pilgrims, were added piecemeal, presumably not following any systematic principle.

Figure 1:

Zentralbibliothek Zürich, Ms. Rh. hist. 27, fol. 31v – codex of fraternisation (www.e-codices.unifr.ch).

The entries in the original layer were written by five hands only and were copied from exemplars that had been sent to Reichenau by the respective monasteries (Autenrieth et al. 1979: LXIII–LXV). The entries were thoroughly planned and follow a clear content structure with regard to the institutional or social belonging of the name bearers. The subsequent entries, on the other hand, show huge variability in layout, list length and technical execution, and are only rarely categorised. Among these entries are lists of several names recorded by a single scribe, but also single entries of one name only. They include copies of exemplars, but may also be ad hoc entries that record a person’s or a group’s visit to the monastery (cf. Section 4.1). In fact, most of the newer name lists are still unexplored, palaeographically, linguistically and historically.

The book was compiled from the start as a “working document” that was intended to grow during later years. For instance, lists of further monasteries were added to the capitula and more names were added to existing lists, which gives some lists the character of proper bookkeeping when, for instance, new names were added continuously as new monks entered a certain monastery. The growing character is also mirrored codicologically in the later addition of more layers (quaterniones) to the original manuscript in around 830 and later in the tenth century.

The manuscript was kept at Reichenau Abbey until the dissolution of the Priory of Reichenau and the Diocese of Konstanz at the beginning of the nineteenth century. After a short stay at Rheinau Abbey, the book came to the Central Library of Zurich where it is stored with the shelf mark Ms. Rh. hist. 27.^[6]

2.2 The liber vitæ of Thorney Abbey

The liber vitæ of Thorney Abbey is contained in the folios at the beginning of London, British Library MS Additional 40,000. The manuscript is an early-tenth-century gospel, with a probable provenance in Northern France (Gneuss and Lapidge 2014: 224–225). The gospels themselves were subject to Old English ink and drypoint glossing from the mid-tenth century onwards (Studer-Joho 2017: 145; Clark 1985: 51), suggesting that the manuscript had moved to England by that date. The palaeographical dating of these hands, however, predates Thorney’s refoundation in 972 as part of the Tenth Century Reformation, and so it would appear that the glosses were added at another location before the book arrived at Thorney. Given Thorney Abbey’s status as one of Bishop Æthelwold’s East Anglian refoundations, it is possible that the manuscript was brought to the monastery at his behest, specifically to provide books for the new institution.^[7] Between c.1090 and c.1200 around 3,200 personal names were added to leaves at the beginning of the gospel (ff. 1r–12v), which according to Gameson (2015b: 115) are “the work of a large number of scribes”.

Thorney’s liber vitæ, then, contrasts with Reichenau’s in that it was not envisaged from the outset as an independent, standalone memorial document; its position in spare or additional folios at the beginning of a gospel marks it as a secondary text.^[8] The very earliest entry, dated by Clark (1985: 51) to the 1060s or 1070s^[9] in what was to become the liber vitæ is an inscription on f. 4r recording a gift to the abbey:

(1)

+ Ælfric wulfwine. Eadgife goldsmiðes geafen to broþer rædenne twegen orn weghenes goldes is on ϸis ilce boc her foruten gewired

[Ælfric and Wulfwine, Eadgifu’s goldsmiths, gave to the brotherhood two oras of weighed gold, which is wired onto the outside of this book.][10]

This entry stands out, as it clearly records the commemoration of the three individuals in exchange for a donation to the abbey, and is reminiscent of other inscriptions detailing the physical history of a book, for example that found in the Codex Aureus (Stockholm, National Library of Sweden, MS A.135). It is possible that this first inscription attracted later names to be added, and so in this respect the entry of the liber vitæ in the gospel had a more organic beginning. However, the earliest layer of true liber vitæ text, on ff 10r–v, is clearly set out for the purpose, as f. 10r begins with the short inscription “Hęc sunt nomina fr(atru)m istius loci”.^[11] Five columns of equal width are ruled across the page, and a number of names were added in the same hand, dated palaeographically to the turn of the eleventh century (see Figure 2). Moreover, as Gameson (2015b: 116) notes, this ruling was also applied to ff. 2, 3, 11 and 12, showing that the project “was undertaken with a degree of foresight, some provision being made for future growth”. It is notable that this arrangement did not last long, as many subsequent entries fail to keep within the narrow columns, and half-way down the second column of f. 10v the arrangement is abandoned in favour of wider, sprawling columns that give way on later folios to long lines. It should also be mentioned that the folios were not completed in the order that they appear in the manuscript, but seem to have been added to in a rather ad hoc fashion, perhaps depending on the amount of space available on a given folio, both in the main writing area, and in the margins, which are frequently utilised.

Figure 2:

© British Library Board. London, BL MS Add. 40,000, f.10r.

Thorney’s name material is not straightforward to interpret. Although the liber vitæ was originally envisaged by its creator(s) as a uniform work with a pre-conceived layout, later entries failed to adhere to the original planned form. There is clear evidence that the earliest entries have been rearranged and copied, presumably from an earlier exemplar. For example, the first entry on f. 10r after the introductory heading is that of Cnut, king of England (1016–35), who is followed by members of his family. The list then proceeds down the social hierarchy, listing archbishops (Æthelnoð, archbishop of Canterbury 1020–38), bishops (Remigius, Bishop of Lincoln 1067–92), and then abbots of Thorney (Godemann, d.1013) (see Figure 3). These ecclesiastical entries end part way down the second column, the remainder of which was originally left blank, and filled with names in a variety of hands at a later date.

Figure 3:

© British Library Board. London, BL MS Add. 40,000, f. 10r.

The third column starts with a group of individuals designated Comes: Turkyll, Hacun, Eoric and Eglaf; the East Scandinavian forms of the names prompted Whitelock (1945: 136) to suggest that this list represents Cnut’s retinue, perhaps representing a visit to the abbey. The latest individuals mentioned in f.10r’s rearranged list are Gunter, Abbot of Thorney (1085–1113) and Robert, Bishop of Lincoln (1093–1123), and by using their life dates this section can therefore be dated between 1093 and 1113. Clearly, the earliest of these individuals were not entered into the list during their lifetimes; the names have been arranged hierarchically and the text composed at a single time, by a single hand. The liber vitæ’s earliest layer of names, then, although written around 1100, contains the names of individuals stretching back over the previous 90 years or so. Case studies 4.2 and 4.3 will focus only on this earliest layer of liber vitæ material.

4.1 Autographs in Reichenau: varying distribution of writing resources

In this case study, we highlight three examples of autograph writing that demonstrate different levels of literacy and show that the manuscript was accessible for writing to a variety of people from various professional and regional backgrounds.

An autograph is usually defined as the hand of the author of a certain work (Long 2014; Overgaauw 2013). In the context of the Reichenau liber vitæ, we classify as autographs entries which include the writer’s name. This may be a single entry or a list with the writer’s name and other names. It is important for the present context that autographs were always produced in the immediate presence of the writer who may have been a member of, or a visitor to, Reichenau Abbey. To identify and describe the autographs, we consider material aspects such as the writing instrument, palaeographic characteristics, the level of technical execution, the layout context of the entry as well as linguistic properties of the names and the historical context.

The first example forms a list of six names in Greek uncials on fol. 40r (see Figure 4):

Figure 4:

Zentralbibliothek Zürich, Ms. Rh. hist. 27, fol. 40r (detail) – entries in Greek capitals (www.e-codices.unifr.ch).

The list is executed by a trained hand and earlier research has stated that the names belong to members of a group lead by Methodius, the archbishop of Pannonia and Moravia, who was kept as a prisoner in a Carolingian monastery, probably Reichenau, during the 870s (cf. Lilie et al. 2013; Zettler 1983). Interestingly, the names ignatius, leo, ioachim and Simon occur also in a sequence on fol. 16r (D3/4) as additions to the list of the living monks from Reichenau. They were all written in Caroline minuscule by one local scribe. The entry methodius is likewise written in Caroline minuscules and by a local scribe at the top of the same list on fol. 15v (A1).^[12] The name SΥΜΕΟΝ occurs elsewhere in rather big and tidily executed Greek capitals on fol. 15r (D3) and there is a third entry in Greek capitals on fol. 55r, ΚΟCΤΑΝΤΙΝΟC that cannot be further identified. The well-done technical execution of all Greek entries indicates that these writers were used to writing. However, there are some open questions with regard to proficiency since both ΛΕΟΝ and CΥΜΕΟΝ/SΥΜΕΟΝ are usually spelled with Omega Ω instead of Omicron O.^[13] The probability is high that the same names in Latin script occurring in the list of the living members of the Reichenau monastery belong to the same persons, and that the local scribe adapted the names to the Latin alphabet when he wrote them down. Hence, he integrated the names into the existing Reichenau list and its linguistic conventions. Zettler (1983: 285, 290) considers the fact that Methodius is added high up in the list to be an indication of his higher status as an archbishop.

While the Greek examples probably give testimony of autographs written by professional scribes from the Eastern part of the medieval Christian world who chose their original writing system to represent their names, another subcorpus with foreign names, the North Germanic names, shows different features. The 743 names probably belong to Nordic pilgrims who travelled by the monastery of Reichenau between the eleventh and twelfth century on their way to Rome and Jerusalem. Most of the names were skillfully written by local scribes in Reichenau and some of them show signs of oral language contact which indicates that the scribes wrote the North Germanic names down by dictation (Jørgensen and Jónsson 1923; Naumann 2009; Waldispühl 2017, 2020). However, three North Germanic names, the male name TOCI, and especially the two female names asa and selva^[14] on fol. 89r (see Figure 5) deviate from the other Nordic entries with regard to their technical execution. The writing is shaky and only partially follows the layout given by the ruled lines. The entries were produced by one or more inexperienced and untrained scribes and that is why we interpret them to be autographs by Nordic pilgrims. The entries themselves are difficult to date palaeographically, however judging by the surrounding entries on the same page they probably originate from the eleventh century. During that time, the most dominant and widely known writing system in the North was the runic alphabet. Hence, the choice of Latin script here might indicate an adaptation of the pilgrims to the script used in the Christian context at the expense of their locally used writing system.

Figure 5:

Zürich, Zentralbibliothek, Ms. Rh. hist. 27, fol. 89r (detail) – autographs by Nordic pilgrims (www.e-codices.unifr.ch).

Our last examples of autograph writing are a couple of instances of scratched entries that were discovered by Andreas Nievergelt and Michelle Waldispühl in October 2019 and reinvestigated in December 2022.^[15] Two of them could be read partly: uuillimuat & liutma… ‘Willimuat et Liutma…(?)’, two names coupled by the et-ligature, in the lower margin to the right on fol. 24r and ‘Burchardus’ in the lower margin to the right on fol. 69v. The other entries could not be decoded at the time. They were probably all scratched into the parchment by a metallic stylus, an instrument common for occasional writing in wax, and can hardly be noticed in normal light. Interestingly, the first entry on fol. 24r has been overwritten by an ink entry which indicates the invisibility of the names at the occasion of the later entry because normally, already-existing names were not overwritten in the manuscript. Dry-point glosses seem to be rare in the Reichenau liber vitæ. In other libri vitæ, only one instance has been discovered so far, in the younger liber vitæ of St. Gall (Erhart 2010; Geuenich et al. 2019). Scratched glosses certainly indicate autographs and suggest even that the name bearers used their personal writing instrument to execute the entry.

Geuenich (Autenrieth et al. 1979: LIX) states that autograph writing in the sense that the named persons wrote their names on their own is “obviously” an exception in the Reichenau manuscript. However, while this is certainly the case for the oldest layer from the 800s, the examples presented here from the younger layers clearly show that autographs are more common in these later times.^[16] As Waldispühl (2019: 213–216) notes, the changes in the conceptualisation and structure of the name lists can be linked to a shift in function and usage of the libri vitæ from books of confraternity between monasteries in the 800–900s to books of general commemorisation in later times. This functional change implies not only a tendency towards the inclusion of more and more names of lay people but supposedly also a wider accessibility of the book for reading and writing. The manuscript might have been located in a place accessible by visitors to the monastery. Moreover, the greater variety in scripting types and situations might mirror the fact that writing skills were more widespread in later times and people were able to make their own record. These observations tie in with Blommaert (2013)’s description of specific sets of resources that are required for writing and that are subject to patterns of social and cultural distribution, of availability and of accessibility. In the younger layers, the manuscript was physically accessible to a wider range of people resulting in entries showing a variety of writers with different technical and graphic repertoires. Moreover, this variability also indicates that the manuscript as a written source was more widely available and that it was intended to include names of people from various backgrounds. The individual level of literacy and personal background was not important either for the writers or for the pragmatic function of the manuscript – the commemorisation of people’s souls.

The choice of writing system and linguistic forms in the lists certainly results from the respective scribes’ graphic and linguistic resources and their community’s conventions. Entries in the Latin instead of the Runic writing system moreover guaranteed that the Reichenau brothers could read the names out loud and include them in their prayers of commemorisation.^[17] However, the use of the Greek alphabet in the autographs presented first could also be interpreted as a distinct choice by these writers to symbolically express their own identity and perform ethnic difference amongst all the other entries written in the Latin alphabet.^[18]

4.2 Choice of script, corrections and language norms

Despite Thorney’s earliest layer being the result of a single scribal effort, the entries on ff. 10r–v display considerable orthographic variation, as might be expected from a text whose source material spans either side of the Norman Conquest. Thorney’s liber vitæ is a witness to the changing personal name trends of the eleventh and twelfth centuries, whereby the original stock of Old English and Scandinavian names was supplemented and subsequently replaced by names with Continental Germanic or Biblical origins. Reflecting this, the names entered on f. 10r–v show a mixture of orthographical and palaeographical traits, and this section will focus in particular on two orthographic variations: <f> or <u> for /v/, and <w> or <ƿ> for /w/. In Figure 6, the writer enters the OE names Ælfwine and Ælfgar following Old English orthographical and palaeographical practices; note the use of graphs specific to Old English, such as initial <Æ>, <ƿ> for /w/, and insular forms of <g> and <r>.

Figure 6:

© British Library Board. London, BL MS Add. 40,000 f.10r: Ælfwine and Ælfgar.

Evidence of Latin or French orthographical norms occurs throughout this earliest layer of liber vitæ material. In Figure 7 a succession of Continental names shows features such as caroline forms of <g> (Hugo) and <r> (the first <r> in Ordricus), and the initial <w> of Walterius rather than <ƿ>.

Figure 7:

© British Library Board. London, BL MS Add. 40,000 f.10r: Radulfus, Hugo, Walterius and Ordricus.

Not every name entry, however, follows the orthographic practices of its etymology, and there is evidence of variation at the palaeographical and orthographical level representing changing orthographic traditions. For example, the name Ælfgifu is made up of two OE elements, (ælf, ‘elf’ and gifu ‘gift’). Traditional OE orthography does not distinguish between voiced and unvoiced fricatives, and so <f> is used for [f] and [v], and <þ> or <ð> for both [θ] and [ð] (Campbell 1959: §50(1)). In the case of ælfgifu, the intervocalic [v] is represented by <f>, and this is the spelling we see in Figure 8. However, the spelling <Ælfgiuu> also appears on the same folio (Figure 9), this time showing the influence of Latin or Anglo-Norman orthographical practices, both of which used <u> to represent the vowel [u] as well as the consonant [v] (Pope 1934: §691).

Figure 8:

© British Library Board. London, BL MS Add. 40,000 f.10r: Ælfgifu.

Figure 9:

© British Library Board. London, BL MS Add. 40,000 f.10r: Ælfgiuu.

Most occurrences of <u> representing [v] occur in female names, however this is largely accounted for by the popularity of names with -gifu as their second element (13 out of 15 female names). Of the names ending in -gifu, only Ælfgifu and Leofgifu appear with <f>, and then only in what appear to be the earliest sections, on f. 10r columns a and c. Male names with <u> for [v] are rarer, most probably because elements with intervocalic [v] are rarer. The one OE instance is Leuing (Leofing), an archbishop, while Scandinavian examples include Toui (x3), Vuah (Ofeig) and Iue (Ivo).

It is not only OE names such as Ælfgiuu which show orthographical forms at odds with their etymologies, however. Examples of insular features appearing in a Continental name are the insular <g> and <r> of Roger le bygod on f. 10v (Figure 10).

Figure 10:

© British Library Board. London, BL MS Add. 40,000 f.10v: Roger le Bygod.

For the scribe of this earliest layer of the liber vitæ, the distribution of <w> and <ƿ> seems dependent on the placement of the graph within the name. The OE name Wulfwine, for example (f. 10r, Figure 11) has initial <w>, but medial <ƿ>. This is broadly consistent with the scribe’s behaviour in this section, where <w> is reserved for initial position (18×), regardless of the name’s etymology (e.g. Wulfstan, Walðef, Wido). Initial <ƿ> occurs only 3 times, each instance a female name (Wulfled, Wulueua). Medial <w>, however, is rare, occurring only once in Godwine on f. 10r column e, along with one instance of the digraph <uu> followed by an unstressed vowel rendered <u> in .^[19] In every other name with medial [w], <ƿ> is used (35×), though apart from Dræwe (Drogo) all these names are OE or Scandinavian. Interestingly, in Figure 11, Wulfwine is split across two lines, yet the second element is written with <ƿ> rather than <w>, maintaining the pattern we have seen of initial <w> and medial <ƿ>.

Figure 11:

© British Library Board. London, BL MS Add. 40,000 f.10r: Wulfwine.

It is clear, too, that some notion of “correctness” was important to the scribe of Thorney’s earliest layer. This could be a matter of factual accuracy, as he distinguishes carefully between individuals with the same name: . (column e) (column e),^[20] confirming that he was indeed intentionally entering these names twice, for two different individuals.

In Figure 12 an error in the text is carefully corrected. When copying, the scribe erroneously attached “et fili(us) ei(us)” [and her son] to Leouiue, whereas it was in fact Leofware who entered into confraternity with Thorney with her son. The error seems to have been noticed immediately, as the passage is deleted, and carefully inserted into the correct position in the text. It is probable that this error was caused through eye-skip, as both individuals have names beginning with <L>; moreover an exemplar which had the names in a similar column-layout, with Leouiue above Leofware, might also have contributed to the original error; such a slip would be an easy mistake to make for a scribe working from an original list formatted in this way.

Figure 12:

© British Library Board. London, BL MS Add. 40,000 f.10r: Leofgifu.

A number of further corrections indicate a desire for correctness, though perhaps stemming from different motivations. In column a Grimgetel is corrected to Grimketel (Figure 13), while in column e Ægelmer is emended to Ægelmær, and Ægeric to Ægelric (Figures 14 and 15).

Figure 13:

© British Library Board. London, BL MS Add. 40,000 f.10r: Grimketel.

Figure 14:

© British Library Board. London, BL MS Add. 40,000 f.10r: Ægelmær.

Figure 15:

© British Library Board. BL MS Add. 40,000 f.10r: Ægelric.

The spelling <Grimgetel> could be accounted for by influence of the initial voiced velar consonant on the second, unvoiced consonant. Bierbaumer (1988: 129–130) suggests that this kind of transmission error arises through the phonological misperception of a consonant. While Bierbaumer’s hypothesis is based on the text in question being dictated, Benskin and Laing (1981: 66) suggest that such errors can also occur as a result of “silent” copying practices:

It may well be that in many such cases what happens is that the scribe moves from copying in a purely visual way to copying via “the mind’s ear”. Instead of reproducing a perhaps laboriously interpreted visual image, the visual image is now interpreted at a glance; and what is held in the mind between looking at the exemplar and writing down the next bit of text, is not the visual symbols, but the spoken words that correspond to them. What the scribe reproduces is then the words that he hears, not the visual images from which they arose: regardless of whether his lips move, he is writing to his own dictation.

Regardless of the origin of the error in copying Grimketel’s name, the important fact is that the scribe noticed the error and corrected it, bringing the name in line with its more usual spelling.

The spellings <Ægelmer> and <Ægeric> are motivated by sound changes and developments in late OE name forms and found sporadically in the Thorney lists ([æ] > [e], and the loss of [l] in ægel- names; see Insley [2015: 270–274]). These two entries in their original form represent names which, by the time of their writing, sounded rather different from their etymological forms. It is possible, then, that the corrections to Ægeric and Ægelmer are motivated by conformity to spelling norms which have their roots in Late West-Saxon, though whether this conformity was motivated through an internalised adherence to such spelling norms or through faithful copying from an earlier exemplar is difficult to tell.

At this point a comparison with the entries on f. 9v is instructive. According to Insley (2015: 277), this folio was the next to be written after f. 10r–v, and contains a mix of conservative and late forms, leading him to suggest a composition date “possibly as late as 1120”. While the scribe responsible for the earliest layer on f. 10r–v uses insular <g> with a fair degree of consistency, to the extent that we might surmise that this was his preferred form of the graph, the scribe of f. 9v only uses Caroline <g>. While f. 10r–v’s scribe uses <ƿ> frequently, but only medially, on f. 9v we find initial <w> mixed with initial <ƿ> (including the form ƿalter), and a more frequent use of <uu> (uuiberti, leofuuini, Ægeluuard). Thus, although both stints appear to incorporate material reaching back well into the pre-Conquest period and stretching up to the time of writing, each scribe demonstrates individual preferences at a graphic level. It is clear that at least some features are down to scribal preference, though it is difficult to judge the role of the exemplars in the compilation of Thorney’s liber vitæ; some corrections point to the text being brought in line with exemplar forms (e.g. Ægelmer > Ægelmær, or efforts to correct eyeskip errors), while others work to “modernise” particular entries (e.g. <haþaƿis> emended to <haþawis> on f.9v). A third group (e.g. grimgetel > grimketel) are clearly corrections of mechanical errors, and are dependent neither on an exemplar for identification, nor on a newer form superseding an older one.

The evidence from Thorney’s earliest layer reveals a mixture of relict forms and those imposed by the scribe, and this evidence takes the form of palaeographical variants alongside orthographic ones. The scribe’s treatment of names ending in the element -gifu reveals a selection of spellings, from the conservative -gifu to forms such as -iue (e.g. Æluiue), however -gifu forms only appear in the first three names with this element; after the Leofgife in column c, all versions of this element use <u> spellings. This could imply that the scribe was working first of all from the earliest sources at hand, and that as he progressed to more modern ones with more modern spellings, his literatim copying practice transmitted the changes in spelling faithfully. Alternatively, he may have copied his earliest sources faithfully, but progressively translated the forms in front of him into his own preferred spellings. The fact that he apparently maintained his preferences for insular <g>, initial <w> and medial <ƿ> throughout his stint does not necessarily mean that he was in all respects a literatim scribe; it is possible that he was a progressive translator orthographically, yet had preferred palaeographical variants which he used more or less consistently, regardless of his exemplar.

4.3 Choice of language

The matrix language of the Thorney liber vitæ is not always consistently or even overtly expressed, as the vast majority of names appear with no other contextualising material. We have seen that the earliest inscription recording the donation of the goldsmiths is in Old English, while the original first section naming King Cnut and his royal party uses a Latin framework. However, if the document was originally envisaged with a Latin framework, this intention was not adhered to; although Latin and Old English are both represented in its earliest layer, neither language choice is maintained throughout the document, and unlike Reichenau there are no further headings organising the entries. Occasional information about the status or accompanying retinue of named individuals is expressed either in Old English or in Latin. Latin examples include Eadwine. , or Ægelsweð. , where Eadwine and Ægelswið are commemorated with family members (wife and son respectively). Examples in OE include Turkyl hoge his ƿif, or Turstan steallere. In these cases, Thurkyl hoge’s wife is remembered alongside him, while Thurstan the staller is distinguished from others bearing the same name by the inclusion of his office.

Example 3, however, makes use of more than one language:

(3)

Eglaf comes. his broðer Vlf.

[Earl Eglaf and his brother Ulf]

The title comes (‘earl’) is written in Latin, but the surrounding text describing Eglaf’s relationship to Ulf is in Old English. Notably these two names are Scandinavian and borne by members of Cnut’s retinue (Whitelock 1945: 134–135), demonstrating that the “ethnicity” of the bearer, the etymology of the name, orthographical tradition, and matrix language may not necessarily align.

Most morphological endings are Latin, though three names contain OE genitives: Asbern haces sunu, Ðorð clapes sunu, and Ælfgifu osgotes wif. The Old English genitives only occur in column c, the first two names in the section comprising Cnut’s retinue, and Ælfgifu in the section following. This distribution would accord well with an underlying exemplar with an Old English framework, as Đorð and another member of the group, Turkyl hoge are both recorded, in English, alongside their wives (“ his ƿif”). Note also that Ælfgifu’s name is one of only three instances with the second element -gifu spelled conservatively with <f>. Nevertheless, the appearance at the head of this colum of four earls (Turkyl, Hacun, Eoric and Eglaf) designated comes, alongside the Walðef comes and comes added above them, poses a problem for this interpretation.

Several names contain Latin abbreviation marks such as for the nominative masculine inflection -us. Of the male names eligible to take this inflection, 56% (84 out of 149 names) contain either an abbreviation or the nominative ending -us.^[21] Name etymology plays a role in the distribution; while only one of the 86 Scandinavian names took a Latin inflection, 39% (32/82) of the Old English ones and 80% (35/44) of the Continental names did (see Figure 16).

Figure 16:

Latin -us endings/abbreviation symbol for -us by name etymology.

Etymology, however, is not the only factor at play. It is notable that of the Old English second elements available, -ric seems to attract -us, whereas the elements -wine, -sig and -wig tend to disfavour -us. Among the names in -ric attracting -us are Godric (6×), Ælfric (4×), Ædric, Ægelric and Wulfric (all 3×); the only -ric not to take -us is Syric (Sigeric). Names disfavouring -us include Ælfwine (5×), Ægelsi (2×), Kynsi, Leofsi, Wulfwi and Ælfwi. It may be that frequency has an effect; the -ric names are all popular ones, in Thorney’s liber vitæ and beyond, while (apart from Ælfwine, which occurs five times in the liber vitæ) the names avoiding -us are less popular.^[22] A further factor governing whether an OE name takes a Latin inflection may be a desire to avoid vowel hiatus. Many names in this category end in -sig and -wig, with a palatalised <g> (Campbell 1959: §428). Abbot Oswius in column a is the one exception to this rule.

Finally, it appears that social status – and especially ecclesiastical status – also has a bearing on which individuals’ names receive Latin inflections. Following the initial list of members of the royal family, columns a and b contain the names of archbishops, bishops, abbots and what has plausibly been identified as members of Thorney’s monastic community in the late-eleventh and early-twelfth centuries. Of the twenty-four OE names listed there, eighteen contain either the abbreviation or – in the case of Oswius – have the inflection written out in full. Four of the six uninflected names – Ædwi, Ægelsi, Ezig and Kynsi – end with the elements -sig or -wig. Notably, the only Old English names attracting an abbreviation outside this list are those ending in -ric (14 instances), suggesting that in this layer of the liber vitæ final element choice is indeed an important factor, along with ecclesiastical status. A precedent for the Latinisation of the names of high-ranking ecclesiastical officials can be found in some charter texts. Seiler (2020: 129) notes that the writers of seventh- and eighth-century Anglo-Saxon charters use script and spelling variation in witness lists to “establish a visual hierarchy among the names of the attesters”. In these documents the names of the king and bishop(s) are in Latinised forms with Latin spelling, whereas the names of witnesses further down the hierarchy are rendered with more Old English characteristics: “if Latin is equated with ‘more formal’, then it makes sense to use the more ‘Latin’ features for the names of the king and bishop and the less ‘Latin’ style for the other, less prominent witnesses” (Seiler 2020: 129). It is possible that the higher rates of Latinisation for this group are down to the names being copied from charter witness lists, however among the monks with OE names only attests a surviving charter.^[23]

Notably, of the thirty-two individuals whose OE names are recorded with a Latin inflection only one, the Abbot Oswius noted above, has the inflection written out in full. This has the effect of rendering the names as “language-neutral units” (Sebba 2012); the frequent use of abbreviation on these personal names pushes the language-specific grammatical information into the background, allowing the reader to access the names in either language (Wright 2002, 2011). From this perspective, then, the symbols in the Thorney list also work as “matters of graphic decorum” (Clark 1992c: 549).

4.4 Dialect adaptations in Reichenau

The final case study deals with the question of dialect choice and the influence the scribe might have had in adapting name forms from other German dialects to the dialect of the scriptorium. As mentioned above in Section 2.1, the oldest layer of the Reichenau liber vitæ from the early ninth century consists of lists from various monasteries on the Continent copied into the codex by five different scribes in Reichenau (Autenrieth et al. 1979: XXIII–XXXIV). This scripting context gives an excellent starting point for the analysis of the scribes’ copying practices of names from other languages and German dialects. Is it possible to see normative attitudes in the Reichenau scribes’ practice? Or do the name spellings reveal a freer handling of dialect features resulting in name forms in Mischsprache? What conclusions about dialect awareness and attitudes can be drawn from these practices? We introduce this section with the presentation of earlier research on these questions, followed by a methodological reflection and a case study of one specific list.

In previous research on the “agency” of the Reichenau scribes with regard to linguistic adaptation, we find contradictory judgements of the same records. In his recent summary of the potential contribution of early medieval Alemannic personal names to linguistics, Geuenich (2018: 2–3) compares variants of the name lemma Rodbert in the name lists from different monasteries copied into the Reichenau liber vitæ. The variants he cites clearly correspond to the regional linguistic realities at the original monasteries (e.g. Hruadpertus for a monk from Alemannic St. Gallen, Ruodbraht from Franconian Fulda, Roppertu from Nanantola in Italy and Frotbertus from Luxeuil-les-Bains in France). Hence, Geuenich arrives at the conclusion that the scribes copied the names accurately from the exemplar lists without adapting them to the phonological and orthographical features of their own dialect. Conversely, in his earlier work on the linguistics of the early medieval personal names from the Fulda monastery, Geuenich (1976: 131, 206–208) excludes the lists of the Fulda brothers recorded in Alemannic Reichenau because they show adaptations to the local dialect (e.g. Ruadbret instead of Ruodbraht) and therefore cannot serve as sources for the Franconian dialect.

This contradictory interpretation of the same records not only shows the complexity of the name material in medieval libri vitæ, but also reveals some methodological pitfalls in variation analysis. In his 2018 article, Geuenich concentrates on the analysis of the spelling variants of one lemma and considers the regional origin of the name forms in different lists. He finds matching patterns because he only considers obvious examples and does not analyse the total of 69 variants for the lemma Rodbert (cf. h 499 hrōth berht, Autenrieth et al. 1979: 109) systematically. These variants also include adapted forms, e.g. the above mentioned Ruadbret in the lists from Fulda. In the current edition, the different name variants are only listed together with their place of occurrence in the manuscript, and possible factors for variation have to be extracted manually for each entry using the facsimile and the palaeographic commentary. Moreover, the macro-perspective analysis of one lemma across different lists that are in many cases written by different scribes as performed in Geuenich (2018) evidently yields other results than the micro-perspective on a small number of lists with names of people from the same monastery as conducted in Geuenich (1976). In the latter case, the regional origin of the name and possible scribal influence are more stable.

Considering these methodological constraints and with regard to the scope of this paper, we select a list from one monastery written by one scribe as an object of study. The oldest list from the Fulda monastery in East Franconia, the so-called “Baugulf-list”, has been chosen to investigate scribal influence on dialect adaptation. The list consists of 458 name entries (Schmid 1978a: 217–218; Schmid 1978b: 572–583) that are distributed over three manuscript pages and arranged in four columns on each page.^[24] The list was recorded by one experienced, local scribe (HA4, Autenrieth et al. 1979: XXXIII, XXIV).

The Old High German Franconian dialect written in Fulda differs in various features from the Alemannic dialect of Reichenau. Most prominently, the diphthongised Germanic *ō (such as in the name element *hrōþ-) already appears as uo in the earliest Franconian texts, whereas in Alemannic, ua was present in the ninth century and only changed to uo in the tenth century (Braune and Reiffenstein 2004: §39a; Baesecke 1928: 104–106). Hence, for the time of the main body of the Reichenau liber vitæ we expect ua-forms, such as -muad- and ruad- for Alemannic and -muod- and ruod- for Franconian.

The names in the “Baugulf-list” from Fulda, however, do not correspond to this rule. Only 15 of the 33 instances with Germanic *ō display the expected Franconian uo-form (e.g. tuoto, muotrat, buoso, hruoduuini, hruodo) whereas 17 appear in the Alemannic form with ua (e.g. hruadfrid, ruadger, ruaduuin, herimuat) and one with u (rudmunt). In the Fuldau necrologies, a contemporary local source from the Fuldau Abbey (cf. Schmid 1978c: 37–95), the names of the same persons show exclusively uo. This clearly indicates that the occurring ua-forms in the “Baugulf-list” are adapted forms that must have been produced in the Reichenau environment. Hence, the scribe certainly played an active role in the copying process. However, the scribe did not adapt the Franconian forms systematically, but shows intra-writer variation. What parameters could explain this variation?

The cognitive linguistic explanation that lexical or semantic transparency of the Franconian names or name elements would lead to the adaptation of elements familiar to the Alemannic scribe can be falsified in several instances. For example, in the names with Germanic *hrōth- as a prototheme, a name element common in both dialects, variation is particularly obvious. Four forms were adapted to hruad- (e.g. hruadfrid, hruadmunt) whereas six forms (e.g. hruodinc, hruoduuini) remained unchanged.

More likely as an explanation is a change in scribal behaviour during the process of copying the list of over 400 names from an exemplar. Considering philological features, the ua- and uo-forms demonstrate two interesting patterns. First, corrections of an o to an a with the same ink colour in two instances (ruaduuin in column B, see Figure 17 and hruadmunt in column D on fol. 31v) suggest that the forms came to attention during the copying process. Second, the ua-spellings are concentrated on the first page (fol. 31v) and the first two columns of the second page (fol. 32r) whereas the uo-forms prevail in the third and fourth column on fol. 32r. On the list’s last manuscript page, the numbers are equal (see Table 1).

Figure 17:

The entry ruaduuin with a correction from an o to an a, Zentralbibliothek Zürich, Ms. Rh. hist. 27, fol. 31v (detail) (www.e-codices.unifr.ch).

Table 1:

The distribution of ua- and uo-forms and different types of mistakes in the “Baugulf-list” from Fulda (Reichenau liber vitæ fols. 31v–32v).

This distribution of ua- and uo-forms on the manuscript pages correlates with different types of errors that indicate a change of copying method in the scripting process (see Table 1). Slips of the ear, such as Iona for Iuno or Liubin for Liubuuin, and ad hoc corrections are both errors that suggest a copying process by dictation. These instances occur mainly on fol. 31v and columns A and B on fol 31r, whereas misreadings from a written exemplar (e.g. Arolf for Asolf) can only be found in the later part of the list.^[25] This distribution might suggest that the scribe noted down the first part of the list by dictation and the second by directly copying from an exemplar. This change of medial circumstance would explain the perception of a diphthong ua instead of uo. This spelling deviation might be due to the pronunciation of the person dictating, or based on the scribe’s spelling according to his own oral repetition, a process known from other instances of scripting by dictation in medieval contexts (cf. Bierbaumer 1988; Waldispühl 2020).

To sum up, the dialectal adaptations from Franconian to Alemannic identified in the Fulda list can certainly be attributed to local scribal influence. However, rather than being indications of linguistic attitudes or conscious normative approaches of the scribe, they are testimonies of oral transmission and the intra-writer variation between adapted and non-adapted forms is likely to show a change in the scripting method from writing by dictation to the copying of visual words during the writing process.