A Carolingian Martinellus: Berlin, ms Phillipps 1877

In the collection of the Staatsbibliothek in Berlin, ms Phillipps 1877 contains a set of writings about St Martin (d. 397), the fourth-century confessor, wonderworker and bishop of Tours.Footnote ¹ The codex belongs to a group of manuscripts known in modern scholarship as ‘Martinelli’, which are today scattered across libraries around the world.Footnote ² The family resemblance of these manuscripts is embodied in a ‘core’ of texts that they usually contain: the hagiographical writings produced by Martin’s contemporary, Sulpicius Severus (consisting of the Life of Martin of Tours, the Dialogues and three letters); a set of verse and prose texts that were inscribed on the walls of various churches in the Touraine (the Sylloge epigraphica); a description of Saint-Martin’s basilica in Tours; a trinitarian formula that is incorrectly attributed to Martin himself in the manuscripts; and some excerpts from the Histories and the Miracles of Martin, composed by Gregory of Tours (d. 594).Footnote ³ There is considerable diversity within the Martinellus family, however: a few of these texts are sometimes omitted, and the order in which they are presented varies. Furthermore, many manuscripts contain additional texts about Martin. These include Paulinus of Périgueux’s verse Life of Martin (c. 460), Venantius Fortunatus’ verse Life of Martin (c. 575), two works about Martin by Alcuin (d. 804), and additional excerpts from Gregory’s writings.Footnote ⁴ A final distinctive feature of early medieval Martinelli is that they contain very little non-Martinian material.Footnote ⁵

ms Phillipps 1877 presents an instructive example. It is one of at least thirty-two Martinellus manuscripts produced before c. 1000 that have survived into the modern era (see Appendix, Table 1).Footnote ⁶ No extant copy dates from before the ninth century (although it is clear, as will be seen, that there were earlier exemplars). The Berlin manuscript is one of at least seventeen Martinelli the production of which can be confidently ascribed to Carolingian Tours.Footnote ⁷ The selection of Martinian materials that it contains is entirely typical, even if there are idiosyncrasies in their presentation. And while the Berlin manuscript is exceptionally lavish, other Tours Martinelli also show signs of being high-status products.Footnote ⁸ What makes ms Phillipps 1877 unusual, however, is the extent of our knowledge about the earliest phase of its life cycle. Although it was copied in the Touraine, the Berlin codex had come to Metz by the later tenth century. In an extra folio added to the front of the manuscript, it is recorded that Bishop Theodoric i of Metz (956–84) presented the work (which he called a ‘book containing the deeds of St Martin’) to the monastery of Saint-Vincent in Metz, his own foundation, as a gift for the eternal benefit of his soul.Footnote ⁹

Following this dedicatory leaf, the ninth-century codex opens with an incipit, lavishly illuminated in gold letters, introducing ‘St Martin’s book about the Trinity’ (that is, the pseudo-Martinian confession).Footnote ¹⁰ The confession is followed by a new title page, introducing Sulpicius Severus’ biography of Martin (fos 3v–35v). After the Life of Martin, we find Sulpicius’ three letters (fos 36r–45r): these describe, respectively, how Martin was miraculously saved from a fire; a vision that Sulpicius received when Martin died; and a description of the confessor’s death while on a visit to the church of Candes. The copyist chose to cut off the third letter – to Sulpicius’ mother-in-law, Bassula – at its midpoint. A decorated title page then announces a new unit of text: ‘Here begins [some writing] about the death of the most glorious St Martin, bishop of the Tourangeaux and confessor’.Footnote ¹¹ The second half of Sulpicius’ letter to Bassula follows (with an ornate illumination, but without a new incipit), describing Martin’s final hours (fos 46r–49v).

A rubricated incipit then introduces a new text within this same section of writings about Martin’s death: ‘Here begins a letter concerning the death of the saintly bishop Martin.’Footnote ¹² This unattributed ‘letter’ is, in fact, a chapter from the Histories, a work produced in the late sixth century by Gregory of Tours.Footnote ¹³ The extract added two useful pieces of information to the saintly dossier: it specified the precise year in which Martin had died, and explained how the Tourangeaux had translated Martin’s body to Tours immediately following his death.Footnote ¹⁴ Another heading then introduces a new text within the same sequence: ‘Likewise [here begins] a letter about the death of the saintly bishop Martin.’Footnote ¹⁵ Once again, what follows is not a letter, but a chapter from Gregory of Tours’s writings – this time from the Miracles of Martin, a four-book hagiographical work. The excerpt (Miracles of Martin i.4) is a spurious story about how Martin’s death was revealed to St Severinus of Cologne (d. c. 404) in a vision.Footnote ¹⁶ Another incipit follows: ‘Likewise a sermon (sermo) of the Blessed Ambrose about the death of St Martin.’Footnote ¹⁷ The extract does not contain any ‘sermon’ by Ambrose: the story that follows is the subsequent chapter from Gregory’s Miracles (Miracles of Martin i.5), describing how Ambrose of Milan (d. 397), too, supposedly received a vision at the moment of Martin’s death.Footnote ¹⁸ The next incipit reads: ‘Likewise another [sermon] [about] when his body was translated.’Footnote ¹⁹ This is Gregory of Tours’s account (Miracles of Martin i.6) of how Bishop Perpetuus of Tours (458/9–488/9) translated Martin’s body from the small church in which Bishop Bricius (397–442) had deposited it to a new basilica.Footnote ²⁰ The scribe rounded off the block of texts about Martin’s death by adding a closing formula that is not found in Gregory’s original work: ‘With the aid of our Lord Jesus Christ who lives and rules with the Father and Holy Ghost through all the ages, Amen.’Footnote ²¹

The next item in ms Phillipps 1877 is the Dialogues, a set of conversations about Martin composed by Sulpicius Severus (fos 57r–127v). It is worth noting here that Sulpicius’ authorship of the Life of Martin, Dialogues and letters are clearly signposted in the incipits, whereas Gregory’s name is entirely omitted from the Berlin codex. Next, we find the Sylloge epigraphica, a relatively stable fixture of the Martinellus tradition (fos 128r–132v). Whereas the first set of verses was produced for Martin’s cell at Marmoutiers, the second set was inscribed on the walls of Saint-Martin’s basilica, the church that Perpetuus had founded in the fifth century to house Martin’s body.Footnote ²² These inscriptions are rounded off with a block of prose text – without any incipit – that notes the dimensions of the basilica, describes its distance from the cathedral, and instructs the reader (addressed in the second-person singular) on how and when to celebrate Martin’s feast days (fo. 133r). The copyists were possibly uncertain whether this short text was another inscription, or had merely been transmitted alongside the epigrams.Footnote ²³

On fo. 133v, we encounter a new full-page incipit, enclosed by decorated arcades: ‘here begins the Life of St Bricius, bishop and confessor’.Footnote ²⁴ This is a biography of Bricius – Martin’s episcopal successor – lifted from Gregory of Tours’s Histories. Footnote ²⁵ Even though the text itself is richly decorated, Gregory’s authorship of the extract is not acknowledged. The Vita is rounded off with a brief closing formula: ‘With the help of our Lord Jesus Christ to whom there is glory, honour, praise, and power alongside the Eternal Father and Holy Ghost for all ages, Amen.’Footnote ²⁶ The Vita Bricii closes the ninth-century manuscript.Footnote ²⁷

Rethinking the Martinelli

ms Phillipps 1877 raises questions consequential not only to our understanding of the Martinellus tradition, but also to broader scholarly debates about the uses of the past in Carolingian Europe. Why did the Tours scribes invest such significant resources into producing the codex? What explains the selection of Martinian writings that were included in the collection? What accounts for the editorial decisions that they made (such as slicing Sulpicius’ letter to Bassula in half to create a sub-dossier of texts about Martin’s death)? And lastly, there is the question of the incipits. Why were extracts from Gregory of Tours’s works introduced with such misleading titles? And why – in Tours, of all places – did the scribes not acknowledge Gregory’s authorship of these excerpts?

The Martinelli are among the most abundant families of manuscripts surviving from Carolingian Europe. The texts that they contain describe Martin’s life, his death, the translations of his corpse and the basilica that Perpetuus built to house his body. These core texts outline events that occurred in the diocese of Tours from the late fourth to the late fifth century. Their authors exclusively belong to the Merovingian (c. 500–751) and pre-Merovingian era. At first blush, then, the Martinelli would appear to have little to say about Carolingian Europe. None the less, this article will suggest that these manuscripts can shed fresh light on the agency of scribes in shaping perceptions of the past in ninth-century Europe. Specifically, it will focus on the presentation of Gregory of Tours’s writings in manuscripts produced at the Tours scriptorium. Although it will argue that most Martinelli were produced for distribution beyond the Touraine, it will close by focusing on one exceptional manuscript – Paris, BnF ms. Lat. 10848 – which was most likely produced for the cathedral of Tours. In this case, the Tours scribes creatively reworked the Martinellus template to serve a much more ambitious project: to create a history of the diocese of Tours starting from its third-century foundation and culminating in the ninth century.

The modern study of the Martinelli has been profoundly shaped by Ernest-Charles Babut’s Saint Martin de Tours (1912).Footnote ²⁸ Babut believed that textual variations within the manuscripts of Sulpicius Severus’ writings could elucidate the early development of Martin’s cult. The Martinelli form only one branch of this early transmission: Sulpicius’ Life of Martin, Dialogues and three letters survive in several early medieval codices without the additional Martinian writings that characterise the Martinellus.Footnote ²⁹ Babut suggested that Martin’s reputation remained contested among the Tours clergy until the time of Bishop Perpetuus, more than sixty years after the saint’s death. He alleged that Perpetuus censored two controversial passages from Sulpicius Severus’ Dialogues in an attempt to heal these divisions. The censorship hypothesis helped to explain why certain manuscripts of the Dialogues omitted those passages.Footnote ³⁰ Alongside editing Sulpicius’ writings, Perpetuus also produced the earliest version of the Martinellus, so Babut believed. This ‘primitive’ Martinellus would have consisted of Sulpicius Severus’ writings and the Sylloge epigraphica, part of which was a short poem that Babut thought had originally functioned as a summary appended to Sulpicius’ writings.Footnote ³¹ He found confirmation for this line of argument in the fact that the Dialogues transmitted in the Martinelli lacked the two controversial passages that he believed Perpetuus had expunged.Footnote ³²

This reconstruction of the early history of Martin’s cult has not won universal approval, however. Already in 1920, Hippolyte Delehaye pointed out that, whilst Martin was clearly a controversial figure in his own lifetime, there is no evidence that the clergy of Tours remained split into two factions up to the mid-fifth century.Footnote ³³ In other words, the censorship hypothesis, so central to Babut’s train of argument, lacks a solid foundation. It is clear that Perpetuus was an early promoter of Martin’s cult, but beyond this circumstantial reasoning there are no grounds for thinking that he prepared the earliest Martinellus.

It is worth stressing at this point that no Martinellus manuscript survives dated to before c. 800. What reason is there for positing a pre-Carolingian Martinellus? It is certain that Sulpicius Severus’ Life of Martin, Dialogues and letters were circulating together by the early sixth century, and indeed it seems that Gregory of Tours himself knew them as a single collection, which he called a historia. Footnote ³⁴ The question, then, is when other Martinian materials were first copied alongside Sulpicius’ works. Babut thought it revealing that the Sylloge epigraphica ostensibly does not include an inscription that the Iberian cleric Martin of Braga produced for Saint-Martin’s around 560, nor the verse inscriptions that Venantius Fortunatus wrote to decorate the walls of Tours cathedral in 589/90.Footnote ³⁵ Neither observation is entirely compelling though. Any of the anonymous poems included in the Sylloge could have been authored by Martin of Braga. Meanwhile, the Sylloge has no inscriptions from the cathedral of Tours. In other words, the absence of Fortunatus’ poems reveals little. In any case, even if the inscriptions were transcribed into manuscript form under Perpetuus, it would not prove the existence of a collection combining the Sylloge with Sulpicius Severus’ oeuvre. As noted above, Babut also believed that the final poem included in the Sylloge was a summary of Sulpicius Severus’ writings. But the location of the poem in the manuscripts, at the end of a set of verses that were inscribed on the walls of Saint-Martin’s, surely indicates that the poem refers to the events from Martin’s life depicted and described on the walls of the basilica itself.Footnote ³⁶ In other words, there is nothing surprising at all about the fact that this poem rounds off the inscriptions. If it were a summary of Sulpicius’ writings, it would presumably feature immediately after one of Sulpicius’ works in the manuscripts.Footnote ³⁷

There is thus no clear evidence that a Martinellus dossier circulated before the time of Gregory of Tours. Indeed, it is worth underlining just how central the latter’s writings were to the Martinellus tradition: all the early medieval Martinelli that I have identified contain excerpts from Gregory of Tours’ writings (see Appendix, Table 1).Footnote ³⁸ By contrast, the Sylloge epigraphica – which Babut considered essential to the ‘primitive’ dossier – is missing from three.Footnote ³⁹ It should also be noted that, while the incipits introducing the excerpts vary, the two blocks of Gregorian writings show signs of originating from a small sample of exemplars. Histories i.48 and Miracles of Martin i.4–6 are almost always sequenced together as a single unit, whilst Histories ii.1 is always presented as a standalone Vita Bricii. Footnote ⁴⁰ These two blocks are often rounded off with closing formulas that can be identified across multiple manuscripts.Footnote ⁴¹ On this logic, it seems likely that Gregory of Tours’s writings formed part of the Martinellus from its very inception. The earliest ‘Martinellus’ dossier would therefore have consisted of Sulpicius Severus’ writings, the five extracts from Gregory of Tours’ corpus, and possibly (though not necessarily) the Sylloge. At some point before about 800, the Pseudo-Martinian confession may have been added into the dossier. Finally, various other texts were added into certain Martinelli during the Carolingian period.

Who produced the first Martinellus dossier, then? It is conceivable that Gregory of Tours himself did so. If he were the compiler, this might explain why his own authorship is so rarely acknowledged in the incipits to the Gregorian extracts contained in the Martinelli.Footnote ⁴² None the less, it would be surprising that Gregory did not mention the compilation of such a dossier anywhere in his extant corpus.Footnote ⁴³ The evidence of the incipits also points against the conclusion that Gregory himself integrated them into the Martinellus. As the case of ms Phillipps 1877 has demonstrated, these incipits inadequately described the contents of the extracts and actively obfuscated their original context within Gregory’s corpus. In fact, one modern cataloguer was led sufficiently astray by the incipits that he listed Miracles of Martin i.5 – Gregory’s account of St Ambrose’s vision – as a text authored by Ambrose.Footnote ⁴⁴ Many medieval readers must have fallen into the same trap. Had Gregory prepared the first Martinellus, we would expect the incipits to signal their position correctly within his corpus, or to correspond to the chapter headings found in the Histories and Miracles of Martin. Footnote ⁴⁵ On this reasoning, the initial compilation of the Martinellus should be viewed as the work of posthumous redactors, rather than Gregory himself.

Hence, we must look instead to the seventh or eighth centuries for the origins of the Martinellus. Indeed, the preparation of the earliest Martinellus could be placed alongside other evidence for the promotion of Martin’s cult at later Merovingian Tours. In the seventh and eighth centuries, the Tours scriptorium was busy producing labels to be affixed to Martinian relics.Footnote ⁴⁶ During the same period, scribes were creatively reworking Gregory of Tours’s corpus. In the first half of the seventh century, a redactor somewhere in Francia produced a shortened recension of the Histories. This ‘B’ recension circulated on a significant scale across the later Merovingian period; one of its exemplars was, in fact, copied at Tours.Footnote ⁴⁷ Finally, in the mid-eighth century, Alan of Farfa (d. 769) compiled a homiliary that included Histories i.48 alongside Sulpicius Severus’ letter to Bassula and an excerpt from the latter’s Vita Martini. Footnote ⁴⁸ It is clear, then, that Sulpicius’ and Gregory’s writings were already circulating together by the mid-eighth century. I would therefore suggest that the compilation of the earliest Martinellus occurred alongside two complementary projects of the period c. 600–750 – the dissemination of the cult of St Martin and the reworking of Gregory of Tours’s corpus.

The production of Martinelli in Carolingian Tours

Most Carolingian Martinelli were written in a ‘Tours hand’. What do we know about copying activities in the ninth-century Touraine? It is clear both from the palaeographical evidence and from dedications that manuscript production occurred on a significant scale at Tours during the period c. 800–53.Footnote ⁴⁹ It is plausible that the cathedral church played a role in copying, but the absence of bishops of Tours in dedicatory notices points against its significance as a centre for book production. It is possible, too, that some writing took place at the monastery of Marmoutiers, Martin’s own foundation. None the less, the invisibility of Marmoutiers in dedications also seems telling.Footnote ⁵⁰ Although scribes with a Tours hand could of course have been working anywhere, the dedications suggest that the abbots of Saint-Martin’s were the primary supervisors of book production in the Touraine. One of the Tours Martinelli, for example, announces its sponsorship by an abbot of Saint-Martin’s: on fo. 183v of Halle, Quedlinburg ms 79, a scribe named Adalbaldus states that he produced the codex at the behest of Fridugisus (abbot of Saint-Martin’s, 804/6–833).Footnote ⁵¹

This copying activity would have occurred in the complex of buildings and churches of which Saint-Martin’s was the focus. This centre was less than a kilometre from the walled area of Tours (within which the cathedral, episcopal residency and a few other churches were situated).Footnote ⁵² During the ninth century, the ecclesiastical status of the basilica was in a process of transition.Footnote ⁵³ In the time of Gregory of Tours, ascetics certainly lived in or around the Martinian complex, but Gregory never referred to Saint-Martin’s as a monasterium (even if he did occasionally refer to an abbot).Footnote ⁵⁴ From the mid-seventh century, charters state that a monasterium formed part of the basilica complex, but a clerical community continued to live there alongside the monks. This uncertainty of status became problematic in the Carolingian period: in 801/2, Charlemagne complained that some inmates of Saint-Martin’s were monks, some were clerics, and others neither monks nor clerics.Footnote ⁵⁵ Over the course of the ninth century, the community of Saint-Martin’s gradually resolved this ambiguity by adopting a canonical lifestyle.Footnote ⁵⁶ Scribal activity probably declined in the second half of that century due to Viking raids, which are recorded in 853, 872 and 903.Footnote ⁵⁷

The abbots of Saint-Martin’s were important actors within the patronage politics of the Carolingian Empire. Alcuin (796–804) was a significant adviser and tutor to Charlemagne. His successor Fridugisus was a witness to Charlemagne’s will and served as arch-chancellor to Emperor Louis the Pious (813–40) from 813 to 833.Footnote ⁵⁸ Under the lay abbot Vivian (843–51), Tours formed part of a group of Loire-valley copying centres tied into the patronage network of Charles the Bald (ruler of West Francia, 843–77).Footnote ⁵⁹ The scribes of Carolingian Tours specialised in the production and export of Bibles – forty-six of which survive.Footnote ⁶⁰ The celebrated ‘Vivian’ Bible, for instance, was dedicated to Charles on behalf of Vivian.Footnote ⁶¹

Who were the Martinelli made for? The sumptuousness of codices such as ms Phillipps 1877 would indicate that the copyists intended them to be distributed to rarefied clienteles beyond the Touraine.Footnote ⁶² It is clear, of course, that the Berlin manuscript was in Metz by the tenth century. Rosamond McKitterick has argued that it displays a script that the Tours copyists reserved for books intended for export.Footnote ⁶³ There was certainly no shortage of demand for writings about Martin: by the dawn of the Carolingian period, evidence for Martinian devotion can be found across Francia, Britain, Iberia and Italy.Footnote ⁶⁴ It is conceivable that Saint-Martin’s produced the Berlin codex for a Carolingian bishop of Metz, such as Charlemagne’s son Drogo, who ruled the diocese from 823 to 855 and acted as Louis the Pious’ arch-chancellor from 834.Footnote ⁶⁵ Equally, the book could have passed via Charles the Bald, given that the latter was crowned emperor in Metz cathedral in 869.Footnote ⁶⁶ It is also indicative that at least fourteen of the surviving Carolingian Martinelli were not copied in Tours script: exemplars must have reached centres as far afield as Auxerre, Freising, Lorsch, Mainz, Regensburg, Saint-Denis and Schönau by the tenth century, if not earlier (see Appendix, Table 1). In sum, the evidence of the Martinelli strengthens the hypothesis that the Tours scriptorium primarily copied books for distribution beyond the Touraine.

Gregory of Tours within the Martinellus

The Martinelli can also shine new light on the reception of Gregory of Tours’s writings in early medieval Europe. Since the 1980s, the manuscript tradition of Gregory’s corpus has been a topic of concerted interest.Footnote ⁶⁷ Reception scholarship on the Histories has proven especially revealing because of the text’s complex transmission history.Footnote ⁶⁸ Although the work must have circulated throughout the early Middle Ages in its original, ten-book form (which Bruno Krusch called the ‘A’ recension), the earliest surviving witness to the complete Histories was copied in the eleventh century.Footnote ⁶⁹ As noted above, a redactor had produced a shorter, six-book version (the ‘B’ recension) within a generation of Gregory’s death.Footnote ⁷⁰ In the ninth century, Carolingian scribes produced a nine-book version of Gregory’s text (the ‘C’ recension) to which they added the fourth book of the seventh-century work known to modern scholarship as ‘the Chronicle of Fredegar’.Footnote ⁷¹ In the early tenth century, a mostly-complete ‘D’ recension was created.Footnote ⁷² These various redactions were attempts to reshape Gregory’s vision of the past to reflect the changing priorities of Merovingian and Carolingian readers.Footnote ⁷³

Recent studies on the reception of Gregory’s writings have overlooked the Martinelli, however.Footnote ⁷⁴ This is surprising, given that Martinellus manuscripts represent one of the most significant transmission pathways for Gregory of Tours’s writings in the early Middle Ages.Footnote ⁷⁵ Just seven ‘B’ manuscripts and nine ‘C’ manuscripts of the Histories survive dating to before c. 1000.Footnote ⁷⁶ By contrast, thirty-one of the Martinelli that survive from the same period preserve extracts from the Gregorian corpus.Footnote ⁷⁷ In particular, the Tours Martinelli offer a unique perspective on how the sixth-century bishop’s writings were understood in his home diocese. Although Gregory’s horizons as an author were in certain respects universal, a sizeable proportion of his writings addressed the history of the Touraine.Footnote ⁷⁸ In the first four books of his Histories, Gregory inserted intermittent notices recording the accession of each new bishop of Tours.Footnote ⁷⁹ He also concluded the Histories with an epilogue containing capsule biographies of his episcopal predecessors, culminating in an autobiographical statement.Footnote ⁸⁰ Gregory rounded off this epilogue by addressing future bishops of Tours directly, urging them to keep his works intact.Footnote ⁸¹ It is likewise relevant that Gregory addressed his Miracles of Martin to ‘the holy lord bishops, sweet brothers, and sons of the Church of Tours’.Footnote ⁸² Given that Gregory presented the local clergy as his legatees, it is especially instructive to consider how the scribes at Saint-Martin’s received and reworked his literary bequest.

Returning to ms Phillipps 1877, it has been noted that readers of that codex would have been entirely unaware that Gregory was the author of any of its contents. The Berlin manuscript is typical in this respect. Of the twenty-six Martinelli produced before c. 1000 whose incipits I have been able to consult, only six feature Gregory of Tours’s name (see Appendix, Table 1). It might be assumed that manuscripts copied at Tours were more likely to acknowledge Gregory’s authorship, but no such correlation presents itself. An explanation for this discrepancy could be that the scribes were copying from already existing compendia of Gregory’s writings, and so were ignorant of the excerpts’ original context. However – as we shall see below – it is very likely that the copyists at Carolingian Tours had access to at least one ‘A’ recension of the Histories and therefore could have known where the extracts had come from.

One motive for deliberately omitting Gregory’s name may have been to give the impression that the latter’s discussions of Martin, Bricius and Perpetuus were near-contemporary with the events described, rather than written some time later. Gregory’s tendentious claims about Martin’s connections with Severinus of Cologne and Ambrose of Milan, for example, would have perhaps appeared more forceful if understood as contemporaneous epistles, like the Sulpician letters. And the credibility of the Vita Bricii might have been undermined had the scribes advertised the fact that it was written around 150 years after Bricius’ death. Whether the silence of the copyists was a cause or a consequence, it is hard to avoid the conclusion that Gregory would not have been well known as a promoter of Martin’s cult in the Carolingian period, even if his writings circulated widely in the Martinelli. It may therefore be necessary to nuance existing claims about Gregory’s influence over conceptions of the past in the early medieval Frankish world.Footnote ⁸³

It is also worth noting the methods that the copyists used to disembed the Gregorian extracts from their original context in his corpus. As seen in the case of ms Phillipps 1877, Gregory’s writings are arranged into two distinct blocks. The first block consisted of one chapter from the Histories (i.48) and three chapters from the Miracles of Martin (i.4–6), which were presented in that manuscript as part of a sequence of texts about Martin’s death (and, implicitly, the posthumous translations of his body). The generic difference between Gregory’s hagiographical and historiographical works was silently elided. Indeed, Gregory may have unwittingly inspired his future redactors when, at Histories i.48, he referred readers to ‘the first book of his miracles’ for further details about Martin’s death.Footnote ⁸⁴ The biography of Bricius (Histories ii.1) forms the second Gregorian block. An attentive reader might have noticed that this unit of text seems to point back to Histories i.48 and includes an allusion to the Miracles of Martin, but the scribe working on the Berlin manuscript actively occluded this connection by separating the two textual units with a mass of other material.Footnote ⁸⁵ The copyists who produced the Martinelli used a set of relatively simple techniques – rubricated incipits and explicits, and short closing formulas – to transform these excerpts into standalone textual units.

Lastly, it is very likely that the Gregorian excerpts found in the Martinelli were put to liturgical use. The sequence of texts about Martin’s death is generally presented as a set of letters (epistolae), sermons (sermones), or little verses (versiculi). These would have been suitable for recitation, probably on Martin’s feast-days (or indeed those of St Severinus of Cologne or St Ambrose of Milan).Footnote ⁸⁶ Meanwhile, the Vita Bricii would certainly have been read aloud on Bricius’ feast-day.Footnote ⁸⁷ The Martinelli are not unique in this respect. Selections of moralising tales from Gregory’s corpus crop up in early medieval homiliary collections.Footnote ⁸⁸ Such scribal efforts to make these extracts into liturgical items present yet another case study in the creative repurposing of Gregory’s writings in the centuries that followed his death.Footnote ⁸⁹

Paris, BnF, ms Lat. 10848: an episcopal past

It has been shown that most Tours Martinelli were produced for export. There is reason to believe, however, that at least one of these manuscripts served an audience closer to home. BnF, ms Lat. 10848 can be confidently dated to the early ninth century (between 816 and 835/7).Footnote ⁹⁰ The Paris manuscript is, like ms Phillipps 1877, a richly decorated book. It also bears certain structural similarities with the Berlin manuscript, which reflect broad commonalities that can be found across the Martinelli, rather than direct dependency.Footnote ⁹¹

The Paris manuscript opens with Sulpicius Severus’ Life of Martin (fos 1r–27r). Sulpicius’ three letters come next (fos 27v–38r). The first of Sulpicius’ letters is split in half, with a single rubricated capital signalling the division between the two sections (fo. 29v). The scribe also divided Sulpicius’ third letter (to Bassula) in half, as had the scribe that produced ms Phillipps 1877: the section detailing Martin’s death is presented on a fresh page with an illuminated capital and its first lines rubricated.Footnote ⁹² Unlike the Berlin copyist, however, the Paris scribe did not shift the second half of the letter into a separate block of writings about Martin’s death. Sulpicius Severus’ Dialogues follow (fos 38v–94r).Footnote ⁹³ The Sylloge epigraphica features next (fos 94r–98r). The description of Saint-Martin’s is separated from the Sylloge with an explicit (fo. 98v), but has no incipit of its own – once again reflecting the ambiguous status of this extract within the tradition. Then we find one of the two ‘censored’ passages from the Dialogues (also without an incipit) that is presented out of sequence in several Martinelli.Footnote ⁹⁴ The pseudo-Martinian De Trinitate appears next (fos 101r–102r).

At this point the first unit of Gregorian materials is introduced. Notably, Gregory is named as the author of these texts: ‘Here begins a sermon of the Lord Bishop Gregory concerning the death of Martin’ (Histories i.48).Footnote ⁹⁵ The Paris copyist and decorators went to considerable lengths in preparing fo. 102v: a spacious incipit alternates between rubricated and non-rubricated capitals, and is followed by a richly foliated ‘A’ (for ‘Arcadius’) at the start of the extract. The second Gregorian extract, Miracles of Martin i.4, is introduced with simple rubrication, indicating that it belongs to the same block of texts about Martin’s death: ‘Likewise another [sermon] by the same person concerning what Bishop Severinus heard upon the death of St Martin.’Footnote ⁹⁶ Miracles of Martin i.5 and i.6 are similarly introduced as sermons authored by Gregory of Tours.Footnote ⁹⁷ Miracles of Martin i.6 here lacks the closing formula that can be found in other Martinelli. Instead, the conclusion to this first unit simply reads, ‘here ends the letter’, recalling the classification of these excerpts as epistolae in some Martinelli.Footnote ⁹⁸ Unlike in ms Phillipps 1877, the Vita Bricii (Histories ii.1) follows immediately, again richly decorated, but (as always) Gregory’s authorship of this excerpt is suppressed.Footnote ⁹⁹ This second unit is then concluded with the same closing formula as in the Berlin manuscript.Footnote ¹⁰⁰

It is the appearance of a third extract from Gregory of Tours’ corpus that sets BnF, ms Lat. 10848 apart from the other Carolingian Martinelli, and accords it a singular role in the transmission history of Gregory of Tours’s corpus. After the Vita Bricii, an austere rubricated heading introduces a text entitled De episcopis Turonicis. Footnote ¹⁰¹ What follows is the epilogue to Gregory of Tours’s Histories (Histories x.31). In this final chapter, Gregory had recorded short biographies of each bishop of Tours up to his own time. In penning these biographical notices, Gregory reprised some information he had provided earlier in the Histories, but added plenty of new material too.Footnote ¹⁰² They address a range of themes: place of birth; year of ordination (dated by regnal years of emperors or of the Frankish kings ruling at the time); the condition of the Christian community in the Touraine; building projects and details of bequests (for which Gregory may have had access to documentation); place of burial; and periods of interregnum.Footnote ¹⁰³

Readers of BnF, ms Lat. 10848 would have found much information relevant to the institutional history, liturgical life and physical environment of Tours in Gregory’s biographies. The manuscript records that the first office-holder, Gatianus, was dispatched from Rome to Tours by an unnamed pope in the time of Emperor Decius (249–51), and suffered persecution at the hands of the city’s pagan inhabitants.Footnote ¹⁰⁴ There was a time gap between Gatianus and the second office-holder, Litorius, who was appointed in 337/8. The readers of BnF, ms Lat. 10848 would have needed to turn back only a few pages to find Gregory’s explanation for the patchiness of the early history of the diocese.Footnote ¹⁰⁵ The notice on Martin referred readers to Sulpicius Severus’ Vita (readily available in the Paris manuscript) and noted the fact that the confessor continued to work posthumous wonders up to Gregory’s time.Footnote ¹⁰⁶ The Bricius mini-biography is largely a summary of Histories ii.1, but with the addition of the saintly bishop’s church-building activities and place of burial.Footnote ¹⁰⁷ Gregory copied Perpetuus’ liturgical calendar verbatim into his biography of the fifth-century bishop, and recorded that Injuriosus (529–46) had introduced the offices of tierce and sext in the cathedral.Footnote ¹⁰⁸ He was also interested in church vessels: Bishop Leo (d. 526) made pyxes, some of which survived up to Gregory’s time. These were presumably recognisable from their inscriptions and perhaps remained in the cathedral into the ninth century. Carolingian readers may also have been interested to read that Leo had been abbot of Saint-Martin’s before his ordination as bishop.Footnote ¹⁰⁹

The culmination of the episcopal succession was Gregory himself. In a brief autobiographical entry, he underlined his own most significant accomplishments: his building projects, his relic-collecting and his writings.Footnote ¹¹⁰ Given the local audience that this concluding chapter seems to envisage, it is hardly surprising that Gregory chose this moment to exhort future bishops to keep his writings intact, impressing upon them their obligation towards the memory of their predecessors.Footnote ¹¹¹ It is relevant, finally, that Gregory used Martin of Tours as one of his chronological orientation-points in the recapitulation of years that closes the epilogue, helping readers to place Martin and the other bishops of Tours in the sweep of universal history.Footnote ¹¹²

Taken together, the epilogue amounted to an episcopal history, not dissimilar to the Liber pontificalis that was composed in sixth-century Rome, and which Gregory may have known.Footnote ¹¹³ The appetite for episcopal histories grew significantly in the Carolingian period: Gesta episcoporum were written for Auxerre, Le Mans, Metz, Naples, Ravenna and Verdun.Footnote ¹¹⁴ These texts represented an expansion of the existing genre of episcopal lists, which continued to be compiled and copied into the Carolingian period.Footnote ¹¹⁵ The authors of Gesta were often cathedral canons – or, in one case, a bishop.Footnote ¹¹⁶ They recorded each office-holder’s circumstances of ordination, building projects, years in office, location of burial, and sometimes charters, letters or synodal decrees cited verbatim.Footnote ¹¹⁷ In some cases, Gesta episcoporum were updated periodically to reflect the succession of further bishops.Footnote ¹¹⁸ By starting their narratives with the first office-holder, who was often a figure said to have been sent from Rome to establish the diocese, the authors of Gesta Episcoporum ensured that the diocesan histories they presented were woven into the broader history of the universal Church. In this way, these texts contributed to a growing sense of institutional self-recognition on the part of ninth-century bishops.Footnote ¹¹⁹

For the readers of BnF, ms Lat. 10848, Gregory’s biographies anchored the bishopric of Tours in a sequence from Gatianus to Gregory. For ease of use, the scribe rubricated and placed in the left-hand margin the number (in Roman numerals) that Gregory had assigned to each bishop, and also rubricated the first letter of each bishop’s name. The final entry is Gregory’s autobiographical notice, and it is accorded special status: exceptionally, Gregory’s number is spelt out in capitalised rubrics as a separate heading.Footnote ¹²⁰ BnF, ms Lat. 10848 would hence have provided a reference volume for anyone wishing to know about the history of the cathedral and its dependent churches. Taken together, Histories x.31, coming at the end of a Gregorian block of texts, complemented the materials relating to Martin, Bricius and Perpetuus contained elsewhere in the manuscript.

The scribe’s particular interest in the history of the diocese of Tours is confirmed by one significant interpolation to the Paris manuscript: after Gregory’s explicit, the same copyist added a two-column list of bishops of Tours down to his own time.

The final entry in the list strongly suggests that the production of BnF, ms Lat. 10848 should be placed in the time of Bishop Landramnus (816–35/7).Footnote ¹²¹ The manuscript was most likely copied at Saint-Martin’s, but the unprecedented emphasis on episcopal succession suggests that the cathedral church – a short walk from the basilica – was the intended destination. A possible analogue can be found in the first instalment of the Gesta pontificum Autissiodorensium (composed between 872 and 875): two of Auxerre’s cathedral canons wrote the Gesta with the help of Heiric, a monk based at the monastery of Saint-Germain, less than half a kilometre from the cathedral church.Footnote ¹²² It is conceivable that a similar collaborative effort between Saint-Martin’s and Landramnus gave rise to BnF, ms Lat. 10848.

While the list contained in the manuscript would patently have provided a useful record for the Tours clergy, it is worth considering whether it also had performative functions. The Libri vitae – the lists of names on whose behalf religious foundations performed prayers, which survive from the mid-eighth century onwards – present a plausible parallel.Footnote ¹²³ Religious foundations activated these books by placing them on the altar and reciting the names during mass. Libri vitae were understood as physical counterparts to the ‘Book of Life’ on which the names of the blessed were thought to be inscribed in Heaven: by integrating the names into their rituals, foundations promoted the well-being of the souls of those listed.Footnote ¹²⁴ Although Libri vitae became a genre with distinct codicological features in the Carolingian period, informal lists can be found in various manuscript contexts: for instance, a copy of a sacramentary from Essen was rounded off with a commemorative list of names.Footnote ¹²⁵ Indeed, other genres of writing could serve comparable purposes: Bishop Dado of Verdun (880–923) wrote a continuation of the episcopal history of his diocese ‘so that the memory of those prelates should be eternally with us – whose names we believe are eternally inscribed in Heaven’.Footnote ¹²⁶ At Tours, the cathedral clergy may well have prayed on behalf of the bishops whose names were listed in the manuscript. This would hardly be surprising, given the transparently liturgical functions of the other texts contained in the manuscript.

Given the complex manuscript transmission of the Histories, it is worth asking how the scribes in Carolingian Tours had access to Histories x.31. After all, the epilogue features in none of the shortened (‘B’ or ‘C’) recensions of the Histories that were produced in the period between Gregory and Landramnus.Footnote ¹²⁷ In fact, BnF, ms Lat. 10848 is the earliest surviving complete witness to Gregory’s concluding chapter.Footnote ¹²⁸ Another early witness was also written in a Tours hand: in the late tenth or early eleventh century, a scribe copied Histories x.31 into a quire now bound into the composite manuscript Universiteitsbibliotheek, Leiden, ms Vossiani latini Q 12.Footnote ¹²⁹ The contents of the Leiden manuscript can in fact shine further light on the study of episcopal history at early medieval Tours. A slightly later scribe copied Histories i.30 (in which Gregory had listed the names of the first bishops of Gaul, including Gatianus of Tours) beneath Histories x.31, indicating a continued concern to investigate the diocesan past.Footnote ¹³⁰ Finally, a fragment containing a list of popes, copied at Tours in the second quarter of the ninth century, was folded into the Leiden manuscript at some point.Footnote ¹³¹ Given that their first bishop had been dispatched from Rome, the Tours clerics could have used this papal list to clarify the early history of their own diocese.

There are minor variations in the wording of Histories x.31 between the Paris and Leiden manuscripts, suggesting the availability of two distinct exemplars for that excerpt in early medieval Tours.Footnote ¹³² Furthermore, three manuscripts produced in tenth-century Iberia contain sequences of Martinian materials that include Gregory’s epilogue in full. These manuscripts are augmented forms of the hagiographical collection that Valerius of Bierzo had compiled in the seventh century.Footnote ¹³³ They distinguish themselves from the Martinelli by containing edifying materials relating to a variety of other saints, by including more extensive materials from Gregory of Tours’s corpus, and by lacking the Sylloge epigraphica.Footnote ¹³⁴ For our purposes, it is significant that the version of Histories x.31 transmitted in the Iberian manuscripts strongly suggests a common source with BnF, ms Lat. 10848, rather than with ms Vossiani latini Q 12 (see Appendix, Table 2).Footnote ¹³⁵ The lost exemplar on which both the Paris manuscript and the Iberian manuscripts based their text of Gregory’s epilogue is presumed to have been made at Tours sometime in the first quarter of the ninth century.Footnote ¹³⁶ The fact that a copy of this exemplar travelled to Iberia further underlines the integration of Saint-Martin’s into regional networks of copying and export.

In short, at least two versions of Histories x.31 were known at Saint-Martin’s in the period between c. 800 and c. 1000: one that formed the model for the Paris manuscript in the early ninth century (Group A), and another that formed the model for the Leiden manuscript at the turn of the tenth century (Group B). Given the lack of independent witnesses to Gregory’s epilogue in early medieval Europe, it is likely that the Tours copyists were working from a full ‘A’ recension of the Histories, either when they made BnF, ms Lat. 10848 or when they produced the earlier exemplar that scholars have hypothesised. By the early eleventh century they had acquired an alternative text of Histories x.31, either in another ‘A’ manuscript of that text or in a codex containing fragments of Gregory’s writings.

In summary, the copyist who produced BnF, ms Lat. 10848 was especially interested in using Gregory’s writings to reconstruct the history of the diocese of Tours. I have posited that this interest stemmed from the fact that the cathedral church was his envisaged audience. By extending the episcopal list up to the ninth century, the scribe joined the city’s Merovingian and pre-Merovingian past to its Carolingian present. This local history flowed from Rome, and hence formed part of a much larger story about the apostolic origins of the Church. This view of the past was far from idiosyncratic: it has been noted that Carolingian episcopal genealogies usually traced their origins to the Petrine city, a device that effectively elevated the diocesan past to the plane of universal Church history.Footnote ¹³⁷ The local and the universal were two sides of the same coin.

In the preceding pages, the scale and diversity of the Martinelli copied in Carolingian Europe have been considered, with a particular focus on two manuscripts written in a Tours hand. Given the unwavering centrality of Gregory of Tours’s writings within the manuscripts, there are strong reasons to question the regnant consensus that there was already a Martinellus in existence before Gregory’s time. Instead, I have proposed that the first Martinellus was produced at Tours in the seventh or eighth century. Its compilation can be placed alongside other contemporaneous efforts to disseminate the cult of St Martin and to creatively rework Gregory’s corpus. With respect to Carolingian Tours, the case study presented here strengthens the existing scholarly consensus that Saint-Martin’s was an unusually productive centre during the ninth century. Most Martinelli were destined for export. ms Phillipps 1877 offers an example of an unusually ornate codex that was given to the cathedral-church of Metz early in its life. The surviving incipits strongly indicate the use of Martinellus manuscripts in the liturgy.

Despite the central place of Gregory of Tours’s writings within the tradition, it is clear that most Martinelli suppress the authorship of the Gregorian extracts. Moreover, chapters from his hagiographical and historiographical corpus are presented side-by-side in these manuscripts, disregarding the narratological features that had distinguished the two genres in the sixth-century bishop’s mind.Footnote ¹³⁸ Given Gregory’s insistence on the integrity of his writings, should the Martinelli be viewed as a violation of his legacy?Footnote ¹³⁹ On the contrary, Gregory himself was an experienced redactor. His method of working was to stitch together self-contained units of text that he had either composed himself or extracted verbatim from existing documents.Footnote ¹⁴⁰ It is also clear that Gregory sometimes hesitated over whether certain stories belonged in his hagiographical or historiographical works.Footnote ¹⁴¹ Gregory’s writings therefore lent themselves to being gutted and recomposed. Had he been able to foresee that future editors would mutilate them, he might have been disappointed, but hardly surprised.

BnF, ms Lat. 10848 presents an exception. In this instance, the copyist went to unprecedented lengths to transform the Martinellus into a living tradition. By adding an up-to-date episcopal list to the epilogue to the Histories, he produced a joined-up history of the diocese of Tours, beginning with Gatianus and ending with the then-current incumbent, Landramnus. Like Carolingian Gesta episcoporum, BnF, ms Lat. 10848 provided an account of the past that was at once local, but placed within a universal story emanating from Rome. Gregory’s authority as a narrator of history was the underpinning of this story. In these terms, it is easy to understand why the scribe who produced the Paris manuscript introduced Gregory with the valorising epithet dominus et episcopus in the incipit to Histories i.48. Indeed, the codicological evidence strongly suggests that Saint-Martin’s had in its archives two copies of the complete Histories by about the year 1000. It was to the Touraine that Gregory had entrusted his writings, and it was here that his memory was best preserved in the centuries that followed.